Changelog for package v4r

1.4.9 (2017-07-28)

merge master into release
fix changelog, package.xml
Merge branch 'fix-rpath-libsvm' into 'master' Fix installation of 3rd party shared libs Closes #29 See merge request !202
Merge branch 'ceres_3rd_party' into 'master' Include and install Ceres as 3rd party See merge request !201
Add CMake command to install libsvm.so
Add 3rd party library directory to the CMAKE_INSTALL_RPATH
fixed old naming issue
install ceres as 3rd party library Because of Ubuntu not fixing a bug in their ceres package [1] we need to add it as a 3rd party lib. [1] https://bugs.launchpad.net/ubuntu/+source/ceres-solver/+bug/1596296
Merge branch 'remove-include' into 'master' Remove useless include See merge request !200
Remove useless include
Update .gitlab-ci.yml
Update .gitlab-ci.yml
update changelog
Test: Enable deb generation on Ubuntu 16.04
merge
Merge branch 'master' into 'master' RGB-D cloud filtering using a sequence of clouds, (batch version, average of depth and colour) See merge request !197
Merge branch 'add_semantic_segmentation' into 'master' add semantic segmentation module + apps and scripts for NYU Datasets See merge request !198
add semantic segmentation module + apps and scripts for NYU Datasets bugfix: renamed one function adapted command line arguments adapted scripts, moved some files from git to repo server Update Readme.md Update Readme.md removed unused include files (esp. openni2_grabber.h) add readme file for semantic segmentation apps update scripts and readme add matlab scripts for nyu depth dataset conversion add apps for semantic_segmentation add module semantic_segmentation
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
changed name of tsf and added example
Merge branch 'master' of rgit.acin.tuwien.ac.at:hannes-prankl/v4r
tsf batch filtering
Merge branch 'fix_global_pipeline_param_bug' into 'master' Fix global pipeline param bug See merge request !196
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'forward_declare_caffe_net' into 'master' Forward declare caffe net See merge request !195
fix bug when global pipeline parameters are not completely defined
Merge branch 'fix_caffe' into 'master' Fix caffe Closes #25 See merge request !194
comment pcl conflict
Merge branch 'forward_declare_caffe_net' of rgit.acin.tuwien.ac.at:t-faeulhammer/v4r
comment definitions
forward declare caffe net avoid caffe include
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
fix missing include paths
Merge branch 'opencv_fix' into 'master' Fix OpenCV Bug for unordered sets See merge request !192
Add Caffe Path hints
Fix OpenCV Bug for unordered sets
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
bug fix: turn off mapping -> tsf-filtering still collected frames
Merge branch 'master' into 'master' Master See merge request !190
bug fix: - TSFGlobalCloudFilteringSimple: set width of the point cloud (if not ec-filtered) - example: parameter setting: do not ec-filter
added modul: camera tracking and mapping (using s temporal smoothing flter - TSF)
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
add multiview keypoint correspondence visualization
Merge branch 'pcl_mesh_renderer' into 'master' Pcl mesh renderer See merge request !189
removed loading of pcl files with pcl::io::loadPLYFile. This method 1. tends to fail and 2. does not exist on the continous integration system.
Cleanup of the changes. Finally got rid of the flipped coordinate system bug.
Added a lot of debug code but principially found the bugs causing the projection to be inverted. TODO: remove debug code!!
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
no idea (must be minor)
Added new constructor accepting pcl meshes.
update doc
update doc
Merge branch 'add_change_detection' into 'master' Add change detection See merge request !188
update doc
update CMakeLists
remove Eval app
fix error when reading floating point occlusion values
avoid running into pcl bug because corr_rejector ransac is not reset
remove unused method
make aligned
lipstick:
read training data despite loading model from file
add more verbosity
allow to load SVM model from file
update default param
update default svm parameter range
fix seg fault when saving model
save trained svm to current working directory
add svm scaling
allow setting kp_transfer
allow to sub-sample views to speed up evaluations
allow to set knn from command line
allow to do icp outside HV
write refined pose
remove condition
remove opencv sift as it is integrated in sift local estimator
weight feature vector so they approximately scale to 1
add some boost program options
remove unused icp parameter
update min dist for cluster factor default parameter value
remove debug
update parameter
temp debug info
remove debugging information again
remove exit
bf
add more debug info
temporary for debugging
update
more info
lipstick:
fix compile error
up
temporary verbose logging
lipstick:
up
update visualization
update params
use openmp for zbuffering
fix min_dist_for_cluster parameter for gc
add multiview visualization for kp transfer
avoid redundant recognition rate computations
use full model for z-buffering in HV
add timing and do some common pre-checks
update author order
lipstick:
pre-compute search radius
set rendering default background-color to white
add missing pieces for remove non-upright objects
serialize hypotheses
add option to remove hypotheses not standing upright on a suppport plane
do not output confusion matrix twice
allow subsampling during rendering
remove unused normal computation for model should be done outside
Merge branch 'fix_normals' into 'master' fix nasty bug in normals computation (flip always towards viewpoint) See merge request !187
use auto
use auto
revert normal referencing
add missing declarations
revert normal referencing
fix visualization bug when skip verification is on
use reference
use reference
fix nasty bug in normals computation (flip always towards viewpoint)
fix doxygen comments
visualize curvature information
fix background color bug
allow to save images to disk
remove leftover for visualiization
add option in xml
just to make sure
add ICRA version of multi-view classifier
add missing piece for param gc grouping
get recognition pipelines
allow to disable correspondence clustering
make deep copy
fix bug when views are empty
add try catch
lipstick:
update params
add keypoint redundancy check
use tanh function for xyz score and re-name parameter
temporarily add some hack to avoid running into bug for outlier cloud visualization
add some more information for timing
fix some timing measurement bug
make shuffle views a parameter
add reading of computation time
lipstick:
fix bug in recognition rate compuation when dot product becomes 1+eps
add multiview evaluation
fix nasty bug in normals computation (flip always towards viewpoint)
add normals visualization
fix multiview and icp stuff, add max views parameter
backup of some old code
remove some deprecated stuff
partially fix multi-view registration
remove unused template parameter
add xtion depth mask
add some logs
add global config
update params
lipstick:
update rec error computation and visualize errors
fix visualization
fix recognition rate bug
add missing files
update dependency list
update color and normals comparison
make HV param private
add dependencies
fix bug when hypothesis does not explain any scene point
lipstick:
remove empty appearance module
adapt L channel
make fitness threshold dependend on visibility
lipstick:
put params in seperate file
lipstick:
Merge remote-tracking branch 'root/master' into add_change_detection
lipstick:
update url for siftgpu
update timings log
re-arrangechecks to hopefully speed-up a bit
fix rec file output and add timings
fix visualization
reset mv
improve performance and brevity of noise model based cloud integration
add copyrights
integrate change detection again
add param for mutliview hv
use seperate parameter class for noise model based registration + fix remaining const madness
add noise model based cloud integration not quite working yet
lipstick:
const madness
update multiview save verification status intrinsically correctly transfer hypotheses across views
add verified member
make objectrecognizer mv compatible
update param
allow online mv rec
opti model - structure (no impl)
1.4.5
Merge branch 'release' of rgit.acin.tuwien.ac.at:markus-bajones/v4r into release
update changelog
set ros distro
update changelog
Contributors: Daniel Wolf, Johann Prankl, Markus Bajones, Sergey Alexandrov, Simon Schreiberhuber, Thomas Faeulhammer, Thomas Fäulhammer, bajo

1.4.8 (2017-04-11)

Merge remote-tracking branch 'upstream/master' into release
Merge branch 'update_siftgpu_url' into 'master' update url for siftgpu Closes #24 See merge request !186
update url for siftgpu
Merge branch 'add_eval' into 'master' Add eval See merge request !185
up
update params
just to make sure
lipstick:
use inner product for normal comparison
up
use pcl_version macro
update param
update z-buffering
use dotproduct for normals comparison
use dist_l2 enum from opencv 2.4
remove highgui include
update params
use vis params
use logarithm
lipstick:
use param
update pose refinement and visible cloud computation
remove color normalization
use pcl stopwatch to allow verbosity setting
use tanh for color comparison
use right color comparison function
fix missing header
reset viewpoint by default to fix visual
small hack to have consistent number of parameters
hopefully fixes compile error
allow inconsistent xml file and just warn
exit when param not okay
use planar filter
use integer for verbosity
initialize depth registration mask
add accidentally deleted header declaration
improve readability
fix wrong sort direction
fix some warnings
use surface normals for z-buffering
rename parameter
update hv parameter
add verbosity
fix
fix compile error
lipstick:
add confusion matrix computation to evaluate global pipeline
reset sensor orientation before segmentation
use distance instead of z component for cut-off
updae eval to allow setting multitoken elements
make remove planes parameter accessible from outside
use distance instead of z component for cut-off
save parameters to file
add initialization
use RGB transform histogram for global color estimator and add parameter class
make histogram row major
use built-in svm cross validation
fix compile error
update global recognizer (mainly add visualization)
update alexnet integration
update global recognizer to allow for multiple feature descriptors per cluster + add visualization option
add to previous commit
use eigen vector instead of PointT type as min max point
add ourcvfh as an option for global concat estimator
use normal estimator from outside
make camera a class variable
remove virtual where not neccessary
add ourcvfh estimator
update global pose estimation and add visualization TODO: check if everything works correctly
update desc
compute discrepancy between centroids of view and centroid of full object model
compute model centroid
fix taking wrong norms (should be sqrt)
use visualization parameter
add parameter for coordinate axis scale
add utility function for aligning two vectors
Merge remote-tracking branch 'root/master' into add_eval
Merge branch 'add-gtest' into 'master' Add GTest See merge request !177
add pose refinement to app
Merge branch 'add_eval' into 'master' Add eval See merge request !182
Merge branch 'fix_iol_runtime_error' into 'master' fix throwing runtime_error See merge request !181
Merge branch 'fix_bug' into 'master' change include order to fix opencv bug Closes #23 See merge request !180
fix throwing runtime_error
fix throwing runtime_error
update default config (elongation can now be part of the feature vector)
update header information
lipstick:
awesome global concate descriptor
remove parameter remnant
lipstick:
fix hist range check
add color to concat descriptor
fix type
add global color descriptor
add concatenation of global descriptor
lipstick:
add a simple global shape estimator
fix visualization bug
remove log copy as glog does not support this apparently
write function to query boost xml value
use glog more frequently
remove volatile remnant
reset normals
lipstick:
make svm param gamma init to 1/num_featurse by default if not otherwise set
add unfiltered global results to vis
make global rec visualization a parameter
fix coordinate system vis issue
fix global recognizers plane alignment
allow to reurn normals from segmentation
update opencv sift
use l1 norm
output recognition rate
revert accidental commit
fix wrong endif for rops
fix sift
fix rops comile error on pcl < 1.7.2
allow to visualize keypoints
update params
allow having multiple keypoint extractor + change default params
update xml
comment try catch
fix bug
tmp commit
performance:
use organized normal computation by default
remove redundant object mask check
fix sift opengl issue TODO: use master thread for sift to use openmp
allow having multiple support radii
lipstick:
remove keypoint cloud and processed cloud from local estimator
Merge remote-tracking branch 'root/master' into add_eval
Merge branch 'add_rotation_check' into 'master' Add rotation check See merge request !179
change include order to fix opencv bug
change include order to fix opencv bug
remove points not belonging to object during initialization
visualize normals in processed cloud
add rops as features not tested yet
make shot support radius a boost paramater
remove nan points before shot computation
remove keypoint indices if the normals are not finite and estimator needs normals
use omp
fix segfault
change default params
allow setting boundary type
avoid keypoint cloud copy if not neccessary
add narf parameters
fix bug
remove config files as they are copied all the time anyway
add parameter for shot and normal estimation
increase normal densitiy in visualization
remove check
change default param
lipstick:
add check if file exists when reading xml files
use stl vector and 💄
lipstick:
lipstick:
lipstick:
make normal computation method a parameter
another try with rootsift
add missing piece
add rootsift as parameter
add l2 normalization
fix
try root sift
revert test
test
fix eval stopping criteria
fix compile error
fix hash
fix compile error on pcl 1.8 (take 2)
fix compile error on pcl 1.8 (take 2)
fix compile error on pcl 1.8
fix compile error on pcl 1.8
add eval
remove eval
add missing executable in cmake
add point types include to gcg to hopefully fix compile error on trusty
add inidividiual hypotheses check with ground-truth data + 💄
Merge branch 'add_param' into 'master' add angular threshold parameter See merge request !176
temp commit
compute normals when init features
allow graph-based correspondence grouping
use cv::Vec3b
also use stl vectors for thresholds
fix not removing already used parameter
make gcg input const
make retrain a parameter for outside
also use 2d stl vector instead of mat for center + fix some warnings
use 2d stl vector instead of cv mat to fix memory leak
fix compile error
Add GTest as a third-party library
add angular threshold parameter
increase model resolution for visualization
make compute recognition rate an apps library
remove plane extractor from global and use segmentation app
wrench:
sparkles:
add angular threshold parameter
Add protection from calling target_include_directories with empty list
Remove INSTALL_TESTS option from CMake
Install only modules marked as PUBLIC
Merge remote-tracking branch 'refs/remotes/root/master' Conflicts: modules/segmentation/src/plane_utils.cpp modules/segmentation/src/segmentation_utils.cpp
Merge branch 'fix_pcl_visualization_issue' into 'master' fix opening multiple pcl visualization windows See merge request !173
Merge branch 'add_simons_plane_extractor' into 'master' add simons plane extractor See merge request !172
Merge branch 'add_hellinger_kernel' into 'master' add hellinger kernel See merge request !171
init params
add boost parameter
remove redundant param init
init params
init params
add boost parameter
add boost parameter
lipstick:
fix compile error
cleanup
fix wrong index
flip normals always towards viewpoint
only vector4f now
temp commit
temp commit
temp commit
tmp commit
temp commit
lipstick:
remove redundancy
reduce redundancy
reduce redundancy
use plane utility functions
use global functions
use const
remove cam
add some plane utils
lipstick:
fix compile error
cleanup
fix wrong index
flip normals always towards viewpoint
only vector4f now
temp commit
temp commit
temp commit
tmp commit
temp commit
lipstick:
remove redundancy
reduce redundancy
use plane utility functions
use global functions
rename config variable name
use const
remove cam
add some plane utils
add seg params
add processed cloud to visualization and use original cloud for hv
update cloud segmenter
allow title change
use boost paths
use boost filesystems paths instead of string will be casted anyway
remove visualization output
extract plane indices and increase performance
fix opening multiple pcl visualization windows define visualizer static
enable plane removal by default
use seperate cloud for processing input
fix opening multiple pcl visualization windows define visualizer static
extract plane indices and increase performance
add hellinger kernel
add simons plane extractor
update recognizer to use new segmentation with explicit plane extraction
add hellinger kernel
add simons plane extractor
Merged branch master into add_evaluation_for_diss
Merge branch 'master' into 'master' ec filter of small clusters - some beautification See merge request !169
ec filter of small clusters - some beautification
Merge branch 'update_io' into 'master' Update io See merge request !167
Merge branch 'master' into 'master' Master See merge request !168
RTMT: remove small clusters before storing the global model
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
use random search for hyperparameter
prepare for local search
fix performance bug
put changes in seperate file and automatically add pairwise changes
add functionality to set output dir
add eval in cmake
do not visualize by default
fix missing occlusion threshold init
fix bug in visualization
init commit
add xml config for multipipeline
lipstick:
add parameter class for multipipeline recognizer
add remove folder
update folder copy
Merge branch 'up_recognition_rate_computation' into 'master' Up recognition rate computation See merge request !164
add chi-square
add parameter for num trees of kdtree in feature matching
use object centroid for translation error computation
make eval compatible to new annotation format
object models are read directly from folder instead from init source
Merge branch 'up_io' into 'master' remove deprecated functions and make paths windows compatible See merge request !165
Merge branch 'revert_disabling_apps' into 'master' revert disabling of some apps and examples See merge request !166
revert disabling of some apps and examples
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
minor: config/ test ImGD-Descriptor
remove deprecated functions and make paths windows compatible
Merge branch 'fix_rf_training' into 'master' bug fix in random forest training See merge request !163
Merge branch 'rec_eval' into 'master' Rec eval See merge request !162
fix pcl version madness
avoid not available getkeypointsindices in PCL < 1.7.2
remove conditional pcl instantiation
lm optimization of the poses (proj. + depth) -- tested
allow to skip verification
Merged branch master into normal_estimation
update keypoint example demo
fix bug when indices are empty
update default param
lipstick:
update normal computations
add z adaptive
init normal class
fix noise model based cloud integration when object mask is not available
fix missing init
update segmentation - use vector<int> instead of pcl::pointindices - seperate plane extraction and segmentation - rename files - move some definitions into .cpp
tmp commit
rename variables
remove siftgpu dependency
remove pcl 1_8 dependency
adds example code for keypoint extraction and filtering
add dependencies
Merge branch 'add_3rd_partry_libraries_as_example_dependencies' into 'master' add dependencies See merge request !159
make segment pcd an app
fix seg fault when input contains nan points
fix wrong if conditions
remove siftgpu dependency
pnp and/or depth ransac version (not tested)
Merge branch 'fix_noise_model_base_integration' into 'master' fix noise model based cloud integration when object mask is not available See merge request !158
add dependencies
fix noise model based cloud integration when object mask is not available
fix redundant typename in keypoint init
Merged branch master into rename_params
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
live version (capture with opencv)
bug fix in random forest training boost random generator is not thread safe, do not access it from within openmp parallel section
update segmentation - use vector<int> instead of pcl::pointindices - seperate plane extraction and segmentation - rename files - move some definitions into .cpp
tmp commit
Merged branch master into rename_params
rename variables
lipstick:
Contributors: Daniel Wolf, Johann Prankl, Markus Bajones, Sergey Alexandrov, Thomas Faeulhammer, Thomas Fäulhammer

1.4.7 (2017-02-23)

Merge branch 'upstream_master' into release
Merge branch 'fix_wrong_sift_init' into 'master' fix missing init See merge request !157
fix missing init
Merge branch 'master' into 'master' added additional parameter for imkRecognizer to specify bin filename See merge request !155
changed imkRecognizer example to handle user filename
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
added additional parameter to set file name for imk-bin-file
Merge branch 'master' into 'master' param change: more accurate flann configuration See merge request !154
param change: more accurate flann
Contributors: Edith Langer, Johann Prankl, Markus Bajones, Thomas Faeulhammer

1.4.6 (2017-02-14)

Merged branch master into release
remove auto keyword as this causes undesired behavior on some machines
Contributors: Thomas Fäulhammer, Unknown

1.4.5 (2017-02-13)

Merge branch 'release' into 'release' Release 1.4.4-internal See merge request !152
Contributors: Markus Bajones

1.4.4 (2017-02-13)

Merge branch 'fixed_missing_parameter_init' into 'master' fix initialize parameters See merge request !151
Merge branch 'ceres-warning' into 'master' Make sure that found Ceres library is a shared object See merge request !150
Make sure that found Ceres library is a shared object Otherwise will get relocation error at linker stage.
fix initialize parameters
Merge branch 'all_updates' into 'master' Segmentation, ML and Keypoint extractor updates See merge request !147
Merge branch 'update_segmentation' into 'master' update segmentation See merge request !146
Merge branch 'ci-ceres-shared' into 'master' Build Ceres in shared mode on CI server See merge request !149
Build Ceres in shared mode on CI server
lipstick:
update classifiers - write a global iniitialization method that allows to initialize any classifier by its type - move parameters outside class to not make them templates
update keypoint extractors - write a global iniitialization method that allows to initialize any keypoint extractor by its type - move parameters outside class to not make them templates
move parameter class
remove segmenter.cpp
update segmentation - write a global iniitialization method that allows to initialize any segmenter by its type - use visualization from utils - move parameters outside class to not make them templates
Merge branch 'new_recognizer' into 'master' New recognizer See merge request !141
put files into v4r module apps to allow to use it as a library fixes alignment issues with potentially wrong library versions of PCL, OpenCV, Eigen, Boost etc
add pointer check
add missing header file to cmake
make classes aligned for fixed-sized vectorizable eigen objects
Merged branch master into new_recognizer
Merge branch 'fix_opencv3_issue' into 'master' fixes #19 (duplicate base type for cv::Feature2D) Closes #19 See merge request !145
fix warnings
Merged branch fix_opencv3_issue into new_recognizer
fixes #19 (duplicate base type for cv::Feature2D) supposed to fix https://rgit.acin.tuwien.ac.at/root/v4r/issues/19
update readme
fix some warnings
do smooth cluster check in global optimization, move visualization option away from param, cleanup
fix visualization
move hv visualization in seperate file/class, update hv evaluation + cleanup
add is_outlier visualization + 💄
fix visualization bug
lipstick:
use auto and fix indentation
make model resolution consistent with scene resolution
fix model color in vis
fix filename
fix compile error
initial commit for mv recognition
fix visualization issue for ghv + fix warning
fix missing camera depth registration mask init for xml constructor
fix hv crash and wrong filename input
lipstick:
read camera parameters from xml file
remove unused recognition files
make path windows compatible
update compute recognition rate
fix error on PCL 1.7.2
update annotation format
fix pcl 1.7 error
use conditional clustering from 3rdParty folder
new recognizer this commit is too huge to comment... sorry a few months work that I can't split up anymore :-P
create class for zBuffering, PCL-2-OpenCV converter and Occlusion Reasoning
use pcl correspondence grouping instead of a copy
Merge branch 'find-devil' into 'master' Look for DevIL library and fail if not found when building SiftGPU Closes #20 See merge request !144
Look for DevIL library and fail if not found when building SiftGPU
Merge branch 'master' into 'master' IMKRecognizer: create model from indices See merge request !142
Merge branch 'cmake-cleanup' into 'master' Cmake cleanup See merge request !143
Remove some occurrences of IOS, WINRT, and ANDROID in CMakeLists
Remove options and config variables inherited from OpenCV
Remove unused CHECK_MODULE CMake macro
Fix defaulting to Release mode
load indices files (additional to to masks) for modelling
added temporal filtering to RTMT-modelling
Merge branch 'update_zBuff_and_OccReasoning' into 'master' Update z buff and occ reasoning See merge request !138
Merge branch 'feat' into 'master' Feat See merge request !137
Merge branch 'master' into 'master' IMKRecognizer: keypoint based, monocular object recognizer See merge request !132
Merge branch 'fix_warning' into 'master' fixes shadowed auto variable warning See merge request !140
Merge branch 'add_db_creation_from_plyfiles' into 'master' creates a recognition database from ply files in folder See merge request !139
removed debug cout
merged
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'fix_some_shadow_warnings' into 'master' fix some warnings (shadowed and unused variables) See merge request !136
Merge branch 'update_ml' into 'master' make svm class better encapsulated See merge request !133
create class for zBuffering, PCL-2-OpenCV converter and Occlusion Reasoning
fixes shadowed auto variable warning
creates a recognition database from ply files in folder (used for new recongition database format)
add pcl visualization parameter class
Merged branch update_ml into master
Merged branch feat into master
Merged branch fix_some_shadow_warnings into master
add shift histogramm function + some 💄
update pcl serialization
update camera class to include depth registration mask
update point cloud properties computation
use new indices
add image crop utility function
some sift updates
fix some warnings (shadowed variables)
Merged branch cleanup_segmentation into master
remove files
Merged branch cleanup_segmentation into master
clean up segmentation and add plane utils removes specific segmenation example from modules
make svm class better encapsulated put files into utility file, add const
add cmake file to find vtk (copied from opencv)
clean up segmentation and add plane utils removes specific segmenation example from modules
Update ObjectDetection.md
make svm class better encapsulated put files into utility file, add const
Merge branch 'update_pcl-1_8' into 'master' put pcl files only availble in PCL 1.8 into 3rdparty PCL folder See merge request !128
Merge branch 'fix_gitignore' into 'master' remove docs from blacklist Closes #16 See merge request !131
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'fix_warnings' into 'master' fix some warnings See merge request !130
remove docs from blacklist
MIT license, colour confidence value, documentation
MIT license header
remove unused pcl trunk files
Merge branch 'fix_opencv3_error' into 'master' fix compile error on OpenCV3 See merge request !129
Fix compile error
remove commment
fix some warnings
fix some more missing pcl_1_8 bits
up
Merge branch 'fix_opencv3_error' into update_pcl-1_8
fix compile error on OpenCV3
use uniform sampling from PCL 1.7 since 1.8 doesn't seem to provide indices
getting mad
up
namespace driving me crazy
add 2d
some more dashes
fix some namespace problems
c++ does not seem to like dashes too much
put pcl files only availble in PCL 1.8 into 3rdparty PCL folder
keypoint based monocular object recognizer
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'update_Cam' into 'master' add fov to camera See merge request !126
remove redundant vertical fiel of view member variable
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'debian_packages' into 'master' create debian packages if commit was tagged. See merge request !110
add fov to camera
Merge branch 'doxy_and_some_minor_beauty' into 'master' Doxy and some minor beauty See merge request !125
Merge branch 'use_constptr' into 'master' use constptr where appropriate See merge request !124
Merge branch 'fix_uniform_sampling' into 'master' make uniform sampling work on PCL versions 1.8+ See merge request !123
some doxygen fixes and 💄
use constptr when neccessary
make uniform sampling work on PCL versions 1.8+
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'some_minor_fixes' into 'master' Some minor fixes See merge request !122
Merge branch 'ceres_fix' into 'master' fix ceres version issue with renamed parameter See merge request !121
add function for principal component analysis (elongation / centroid)
neglect .autosave files
fix quaternion computation in case not normalized
add focal length to boost program options
remove openni include
fix ceres version issue with renamed parameter
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'feat1' into 'master' Feat1 This adds some functions used for recognition / hypotheses verification See merge request !118
Merge branch 'add_serialization' into 'master' Add boost serialization for common pcl types See merge request !117
Merge branch 'add_doxygen_config' into 'master' add doxygen config file to be able to generate Doxygen based documentation See merge request !116
Merge branch 'use_pcl_eigen' into 'master' remove redundant copy of PCL file cleans up v4r repository (redundant PCL copy - PCL is a required dependency anyway) See merge request !115
Merged branch master into debian_packages
add computeMaskFromImageMap function
minor code reduction
add initial histogram equalizer not tested yet
add colorcomparison method enum
remove unused std_msgs serialization
add boost serialization for common pcl point cloud types
some more redundant pcl copies replaced
add doxygen config file to be able to generate Doxygen based documentation
fix namespace issue
replace v4r/common/eigen.h with pcl/eigen.h
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'warningfix_pcl' into 'master' Fixed warnings in PCL So, this should fix every warning from external libraries. I did this by adding the necessary SYSTEM parameter to the according "(target_)include_directories" calls. As far as i see it it does not suppress any warnings in our own headers (as intended) but due to my lack of understanding of CMAKE and our V4R structure i suggest somebody might want to test it. See merge request !113
Fixed another mistake... credits goes to Sergey.
Fixed missing }
removed comments.
Being more selective at suppressing warnings. Being more selective at suppressing warnings as now only warnings from external header files will be suppressed.
Merge branch 'warningfix_keypoints' into 'master' Fixed some Warnings in Keypoints As the title says. I hope i broke no real functionality. See merge request !100
Update V4RUtils.cmake
Suppress warnings in external headers. Added the SYSTEM property to V4RUtils.cmake where it is needed to suppress pcl warnings of external include files.
fix to lowercase path name
rename header files to be included in the packaging process
update changelog
Merged branch master into debian_packages
Merge remote-tracking branch 'v4r-master/master'
Update .gitlab-ci.yml Install sed, use it to compile on 8 cores
Update .gitlab-ci.yml Make sure v4r is installed in /usr
Merged branch master into master
Merge branch 'attention' into 'master' Integrate attention based segmentation Now works on Ubuntu 14.04 and 16.04 @michael-zillich-1 @msuchi Can I get some feedback before I merge this? See merge request !112
disable deb packages on xenial for now. issue with libceres to blame
Adaptions for v4r source code for compiling under Ubuntu 16.04. + changed include from "cvmath" to <cvmath> + changed namespace of isnan to std::isnan this is tested for opencv 2.4 and pcl 1.7.2 which have to be set when launchi9ng cmake: cmake -DPCL_DIR=<path to pcl> -DOpenCV_DIR=<oath to opencv 2.4>
Merged origin/attention_segmentation into master
Merge branch 'Update_Contributing.md' into 'master' Update contributing.md Added usage of setup.sh to CONTRIBUTING.md See merge request !111
need to install devscripts
need to install python-bloom before we can use it
need to install python-bloom before we can use it
Update CONTRIBUTING.md
update gitlab-ci file
Added dependency installation "how to" to Contributing.md
create debian packages if commit was tagged.
Merged branch master into master
Merge branch 'ubuntu16.04+opencv3' into 'master' Ubuntu16.04+opencv3 @ghalmets See merge request !103
pass 2 parameters to setup. ubuntu and ros codename
Removed some warnings in keypoints.
escape variables
update setup.sh and gitlab-ci.yml to be more generic
gitlab's lint checker says it is OK now. Let's see.
fix gitlab syntax after strange merge issue
Merge remote-tracking branch 'refs/remotes/upstream/master' Conflicts: .gitlab-ci.yml
Merged branch master into master
Merge branch 'bug_transposed_rendering' into 'master' Changed camera matrix input and output of the pointcloud generation class. Now i… This is the bugfix to: https://rgit.acin.tuwien.ac.at/root/v4r/issues/13 See merge request !98
Merge branch 'Install_dependencies' into 'master' Adding Setup.sh Added Setup.sh to v4r for a more convenient dependency installation. .gitlab-ci.yml was edited to use setup.sh to keep the script in the CI loop. Workflow: git clone git@my-awesome-v4r-repo cd v4r ./setup.sh mkdir build && cd build cmake .. make -j8 See merge request !96
Update Readme.md
Update Readme.md to rgit and added usage of ./setup.sh
Changed camera matrix input and output of the pointcloud generation class. Now it is not transposed, or does not has to be transposed anymore.
Update .gitlab-ci.yml
Added Setup.sh
Add script for first build Installing rosdep and dependencies, building v4r.
update apps to include all programs for attention based segmentation
update attention_segmentation module
Add first sample app for attention based segmentation! Yes it works.
fix header files
fix cmake mistake
add and activate opennurbs and on_nurbs
shifted around attention_* files
Merged branch master into master
fix cmake file
add opennurbs as build option for V4R
reflect change of opennurbs directory
add cmake find file for openNurbs
moved opennurbs to 3rdparty
delete autosave file
change dependencies
bring in all the files for attention segmentation -HACK
Update examples after eputils merge
Small changes because of the eputils merge into attention_segmentation
We no longer need/have a v4r_eputils module
move eputils into attention_segmentation
Add examples for attention based segmentation
Adapt to new v4r structure for attention based segmentation mainly namespace changes, V4R_EXPORTS, etc.
Add missing files for eputils
Adapt to new v4r structure namespaces, V4R_EXPORTS, etc.
small changes to bring attention_segmentation into the new v4r structure
small changes to bring eputils into the new v4r structure
Inital copy of attention_segmentation from v4r svn
Inital copy of eputils from v4r svn
Merge branch 'master' into 'master' v4r now compiles with OpenCV 2.x and 3.1 on Ubuntu 14.04 and 16.04 See merge request !94
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
v4r now compiles with OpenCV 2.x and 3.1 on Ubuntu 14.04 and 16.04
Merge branch 'master' into 'master' temporal filter (tracks pose and integrates several rgb-d frames) incl. bug fix: default param in common/occlusion_reasoning.cpp See merge request !93
Merge branch 'master' into 'master' Update of Contribution.md with results from Structure Workshop. I have merged the minutes of V4R structure workshop into the Contribution.md See merge request !92
Update CONTRIBUTING.md
Fixed some Typos
Update CONTRIBUTING.md
temporal filter (tracks pose and integrates several rgb-d frames)
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Update CONTRIBUTING.md Update description v4r exports
Update CONTRIBUTING.md minor changes
Update CONTRIBUTING.md: Formating
Update CONTRIBUTING.md: added sections: "Structure", "Documentation", and "How to Build V4R?".
Manually set PCL_APPS_LIBRARY
Merged branch ubuntu16.04+opencv3 into ubuntu16.04+opencv3
I think this is it.
say yes to apt-get. all the time
No debug symbols installed
build ceres from source
We can force the dpkg installation
handle install with apt-get force
Next Ubuntu hack
fix stupid Ubuntu typo
Introduce hack because Ubuntu
rosdep really needs sudo. install it.
No sudo in xenial image
No sudo in xenial image
fix ubuntu version. and again.
fix ubuntu version
Also build on Ubuntu 16.04
fix: no default values
I think this is it.
say yes to apt-get. all the time
No debug symbols installed
build ceres from source
We can force the dpkg installation
handle install with apt-get force
Next Ubuntu hack
fix stupid Ubuntu typo
Introduce hack because Ubuntu
rosdep really needs sudo. install it.
No sudo in xenial image
No sudo in xenial image
fix ubuntu version. and again.
fix ubuntu version
Also build on Ubuntu 16.04
Merge branch 'Test' into 'master' Update AUTHORS See merge request !89
Merge branch 'libsvm' into 'master' fix libsvm dependency in package.xml See merge request !90
fix libsvm dependency in package.xml
Update AUTHORS
Merge branch 'add_boost_dependency' into 'master' Update hypotheses_verification.cpp (wtf? I just commited this change) See merge request !88
Update hypotheses_verification.cpp
Update hypotheses_verification.cpp (wtf? I just commited this change)
Merge branch 'master' into 'master' Master See merge request !82
Merge branch 'add_boost_dependency' into 'master' add missing boost dependency See merge request !86
add missing boost dependency
Go back to use standard ubuntu trusty docker image This is easier to support in the future.
Merged branch master into master
Merged branch master into master
Merge branch 'fix_vector_type' into 'master' Fix vector type See merge request !79
Update hypotheses_verification.cpp
Update hypotheses_verification.cpp
fix vector type
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
fix vector type for new histogram interface
Update Readme.md
Merge branch 'master' into 'master' Use docker image that has those dependencies already installed See merge request !78
Use docker image that has those dependencies already installed
Merge branch 'master' into 'master' Update .gitlab-ci.yml See merge request !77
Update .gitlab-ci.yml
Merged branch master into master
Update .gitlab-ci.yml
Update .gitlab-ci.yml
Merge branch 'some_fixes' into 'master' Some fixes See merge request !56
Update .gitlab-ci.yml
Merged branch master into master
Update .gitlab-ci.yml yet again. ;-)
Update .gitlab-ci.yml
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
put some header defintions into cpp files and remove .hpp files
Update .gitlab-ci.yml
put miscellaneous functions into more specific files
Update v4r_style_guide.md
fix merge conflict
Merged branch master into master
added: only small inline functions
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Merged branch master into master
added: keep pull requests short
fixed typo
fixed typo
clean up .gitlab-ci.yml
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
add depdendencies description
Update .gitlab-ci.yml
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
add contributing and style_format doc files
Update package.xml test if this compiles now
Update .gitlab-ci.yml Continue on rosdep errors. Arrrrrr
Update .gitlab-ci.yml fix syntax
Update .gitlab-ci.yml specify the ROS version (needed to resolve packages from package.xml)
Update .gitlab-ci.yml We need wget as well.
Update .gitlab-ci.yml
Update .gitlab-ci.yml Seems like we need cmake after all
Add .gitlab-ci.yml First try
update color transformation and comparison
use boost::dynamic_bitset instead of vector<bool>, add camera class, put some definitions into header files + some code refactoring
fix roi when reaching boundary
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'master' into 'master' Ubuntu 16.04 compatibility See merge request !54
Merge pull request #67 from strands-project/ubuntu1604_compatibility Ubuntu1604 compatibility
Merge remote-tracking branch 'hannes/master'
add pcl time header
Merge remote-tracking branch 'v4r-master/master'
add bounding box function
seperate definitions from some header files to reduce compile time
fix some warnings
Merge pull request #66 from strands-project/sync_from_gitlab Sync from gitlab
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'new_try' into 'master' New try See merge request !53
Merge pull request #64 from strands-project/new_try [WIP] New try
add timing for pose refinement
update citation file
put pcl_opencv functions from header into implementation file
some code optimization
some changes for compiling with Ubuntu 16.04
some beauty
add script for obtaining alexNet CNN
use const
make destructors virtual for virtual classes
remove empty file
remove broken files
put test data into directory
fix existing directory warning
add docs for recognition update get_TUW script
add script for downloading 3dnet test data
add doc for RTMT
include scripts to obtain training data from TUW and 3dNet
include missing mean substraction in alexnet feature estimation
update for shape cnn classifier to work
fix wrong model assembly resolution
fix compilation errors for eval and app pieces
fix missing clear of indices when no keypoints are detected
remove voxelgriddistancetransform method call
remove default typename in createIndicesFromMask method to allow usage without c++11
add global hypotheses non-maxima surpression to rejection method
group hypotheses by segmentation cluster
add online averaging function
add hyp stuff (should have been staged earlier)
remove EDT stuff
check if all views are trained during initialization (not just if directory exist)
put boost program options directly into parameter classes, merge ghv with hypotheses verification class
make seperate table class
minor fixes for save pose into pcd
update some visualization functions in recognition
remove sift based alignment in incremental object learning class
use new segmentation class and provide combined feature, keypoint and segmentation layer
hopefully fixing Caffe optional dependency
up pcl version
fix compilation error caused by addcoordinatesystem if used with PCL < 1.7.2
add esf classifier again
fix typo in openmp call
fix some warnings
fix bug in optional dependening on caffe and rendering
change default params and do not instantiate harris and iss keypoint extractor on PCL versions < 1.7.2 (as keypoint indices is not available for these versions)
make recognition library dependency to rendering and Caffe optional
move some hpps into cpps
skip recongition rate computation if output file already exists
add nicer visualization for paper
add todo comment
add eval for rec rate over occlusion
fix crop bug in pcl opencv conversion
fix min fitness threshold scaling
flip table plane towards viewpoint and make parameter for min points accessible
make resolution as an integer in mm
add coordinate system for visualizing recognition results
fix bug in color conversion
change default parameter for svm cross validation
make smooth segmentation parameter scale with depth
avoid table plane filtering for initialization in local recognizer
add parameter options for smooth clustering
add dense SIFT option (not tested yet and only available for SIFTGPU)
add smooth clustering and linear thresholding of model fitness threshold (with visibliity)
use multi-plane segmentation for local recognizer to find heighest table plane
fix visualization for recognition rate computation when cloud sensor header is changed
temporary remove parallel call of recognizer QH6205 qhull error (qh_initqhull_start): qh_qh already defined. Call qh_save_qhull() first
fix bug in compute recognition rate
ignore multiview and esf object classifier for now
make model fitness threshold adaptive to visible ratio (TODO: write a proper function)
use bigger rendering points in model cues visualization
fix wrong sigma for AB color components
remove table plane debug visualization
rename some recognition init parameters
reset view to do not mess up visualization in evaluation recognition example
add option to just compute dominant plane (without clustering)
fix bug with multiple call to recognize if recognizer is local
add all the options for initialization
make local recognizer more modular
fix bug in knn_sift initialization
add missing iostream include in classifier.h
add opencv sift option again (NOTE: Not tested yet)
remove keypoint rounding stuff in sift
rewrite local estimator interfaces
remove redundant files, take into acccount sign ambiguity of eigen vectors for global recognizer
fix bug with missing normal computation
migrated feature estimator changes (except eigen matrix). kinda working but only for first test view it seems
add global recognizer
add ourcvfh pcl trunk version, fix view all point clouds in folder
merging svmwrapper, classifier classes, keypoint extractors... still working
merged many things from rec_optimization_isolated branch (hyp generation still working - verificaiton not)
add point cloud to eigen matrix conversion with indices
add ptr for gcg
remove old comments
add ptr for gcg
fix merge conflict
add vector sort with indices return
add some histogram functions
add cielab2rgb transformation
some code polish in graph based geometric consistency grouping
avoid some warnings
add visualization of model keypoints
fix visualization of correspondences
remove global requirement for samples to have all modules enabled it now only looks for the individual dependency of each sample and compiles just the ones which meet their dependencies
using parameter class for gcg when gcg is used... small code polish
addition to previous commit
fix color conversion compilation error in case of PointXYZ instantiation
make ghv compile for PointXYZ type instantation as well
fix error with color retrieval in verification code
optimize speed
fix bug in model assembly
remove parameter output
speed up verification evaluation by compressing scene fitting matrix
add recognition rate evaluation
make it compile for PointXYZ as well
compute visible model cloud on multiple views
merge hv_go3d into ghv (not ready yet) optimize visible model cloud computation in verification
add depth image computation in zBuffering class (remove XYZRGBA template instantition)
split code into more functions, add omp sections again, and some minor beauty
add replace moves again by checking pairwise intersection matrix
enhance pairwise intersection computation by fixing smoothing, speeding up computation and adding erosion
remove some more obsolete code
use new verification cost function and remove obsolete code pieces
add smoothing function to zbuffering (does not work properly though)
add function to remove column or row from eigen matrix
fix compiler error in change detection module
implement pairwise intersection computation in verification algorithm
add rendering function in zbuffering (explicit)
use local point color to compare color
delete obsolet member variables
delete count active hypotheses function in verificitation (as it is not used anyway)
make update function use member variables instead of having to pass them as an argument
do not use weights for outliers - just ratio of number of outliers compared to visible points
clip noise model based radius for inliers search
rename variable and do label check earlier to avoid redundant processing
fix seg fault when not using icp for pose refinement
reset camera view point in object recognizer to avoid messing up visualization
clip max viewing angle in noise model to 80 degrees to avoid huge noise terms (was 85)
use noise model for model explained points
fix ignore color even if exists check
fix wrong use of row and column counter in self zbuffering module
do incremental smooth clustering via noise model (not finished yet)
make visualize go cues a switch parameter
add a static function to query noise level for a single point
temp commit
Integration of change detection into recognition module
Annotation of changes in GT data
Change detection module added
Compilation fix: duplicated pragma
add merge for multiview
normalize optimization variables
fix multipipeline merging of hypotheses when disabled. Also skip merging of ident hypothesis
fixed self occlusion reasoning
add pose refinement fix noise model based cloud integration for just one input cloud as well as for no indices given
fixed points on plane side variable in ghv
working again
explained and unexplained points seem okay
fix merge conflict
fix merge conflict
Merge pull request #63 from taketwo/remove-x86 Remove all mentions of x86 and x86_64 in CMake scripts
Remove all mentions of x86 and x86_64 in CMake scripts
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
use object indices also for unfiltered registered model cloud and only save filtered input clouds if debug option is set
Merge branch 'master' into 'master' added quick fixed to handle some range check exceptions needs proper handling soon See merge request !51
1.3.3
1.3.2
Merge remote-tracking branch 'upstream/master'
add missing Xxf86vm lib
Merge remote-tracking branch 'remotes/upstream/recognition_update'
Contributors: Georg, Georg Halmetschlager-Funek, Johann Prankl, Markus Bajones, Markus Suchi, Martin Velas, Michael Zillich, Sergey Alexandrov, Simon Schreiberhuber, Thomas Faeulhammer, Thomas Fäulhammer
Merge branch 'attention' into 'master' Integrate attention based segmentation Now works on Ubuntu 14.04 and 16.04 @michael-zillich-1 @msuchi Can I get some feedback before I merge this? See merge request !112
Adaptions for v4r source code for compiling under Ubuntu 16.04. + changed include from "cvmath" to <cvmath> + changed namespace of isnan to std::isnan this is tested for opencv 2.4 and pcl 1.7.2 which have to be set when launchi9ng cmake: cmake -DPCL_DIR=<path to pcl> -DOpenCV_DIR=<oath to opencv 2.4>
Merged origin/attention_segmentation into master
Merge branch 'Update_Contributing.md' into 'master' Update contributing.md Added usage of setup.sh to CONTRIBUTING.md See merge request !111
Update CONTRIBUTING.md
Added dependency installation "how to" to Contributing.md
Merged branch master into master
Merge branch 'ubuntu16.04+opencv3' into 'master' Ubuntu16.04+opencv3 @ghalmets See merge request !103
pass 2 parameters to setup. ubuntu and ros codename
escape variables
update setup.sh and gitlab-ci.yml to be more generic
gitlab's lint checker says it is OK now. Let's see.
fix gitlab syntax after strange merge issue
Merge remote-tracking branch 'refs/remotes/upstream/master' Conflicts: .gitlab-ci.yml
Merge branch 'bug_transposed_rendering' into 'master' Changed camera matrix input and output of the pointcloud generation class. Now i… This is the bugfix to: https://rgit.acin.tuwien.ac.at/root/v4r/issues/13 See merge request !98
Merge branch 'Install_dependencies' into 'master' Adding Setup.sh Added Setup.sh to v4r for a more convenient dependency installation. .gitlab-ci.yml was edited to use setup.sh to keep the script in the CI loop. Workflow: git clone git@my-awesome-v4r-repo cd v4r ./setup.sh mkdir build && cd build cmake .. make -j8 See merge request !96
Update Readme.md
Update Readme.md to rgit and added usage of ./setup.sh
Changed camera matrix input and output of the pointcloud generation class. Now it is not transposed, or does not has to be transposed anymore.
Update .gitlab-ci.yml
Added Setup.sh
Add script for first build Installing rosdep and dependencies, building v4r.
update apps to include all programs for attention based segmentation
update attention_segmentation module
Add first sample app for attention based segmentation! Yes it works.
fix header files
fix cmake mistake
add and activate opennurbs and on_nurbs
shifted around attention_* files
Merged branch master into master
fix cmake file
add opennurbs as build option for V4R
reflect change of opennurbs directory
add cmake find file for openNurbs
moved opennurbs to 3rdparty
delete autosave file
change dependencies
bring in all the files for attention segmentation -HACK
Update examples after eputils merge
Small changes because of the eputils merge into attention_segmentation
We no longer need/have a v4r_eputils module
move eputils into attention_segmentation
Add examples for attention based segmentation
Adapt to new v4r structure for attention based segmentation mainly namespace changes, V4R_EXPORTS, etc.
Add missing files for eputils
Adapt to new v4r structure namespaces, V4R_EXPORTS, etc.
small changes to bring attention_segmentation into the new v4r structure
small changes to bring eputils into the new v4r structure
Inital copy of attention_segmentation from v4r svn
Inital copy of eputils from v4r svn
Merge branch 'master' into 'master' v4r now compiles with OpenCV 2.x and 3.1 on Ubuntu 14.04 and 16.04 See merge request !94
v4r now compiles with OpenCV 2.x and 3.1 on Ubuntu 14.04 and 16.04
Merge branch 'master' into 'master' temporal filter (tracks pose and integrates several rgb-d frames) incl. bug fix: default param in common/occlusion_reasoning.cpp See merge request !93
Merge branch 'master' into 'master' Update of Contribution.md with results from Structure Workshop. I have merged the minutes of V4R structure workshop into the Contribution.md See merge request !92
Update CONTRIBUTING.md
Fixed some Typos
Update CONTRIBUTING.md
temporal filter (tracks pose and integrates several rgb-d frames)
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Update CONTRIBUTING.md Update description v4r exports
Update CONTRIBUTING.md minor changes
Update CONTRIBUTING.md: Formating
Update CONTRIBUTING.md: added sections: "Structure", "Documentation", and "How to Build V4R?".
Manually set PCL_APPS_LIBRARY
Merged branch ubuntu16.04+opencv3 into ubuntu16.04+opencv3
I think this is it.
say yes to apt-get. all the time
No debug symbols installed
build ceres from source
We can force the dpkg installation
handle install with apt-get force
Next Ubuntu hack
fix stupid Ubuntu typo
Introduce hack because Ubuntu
rosdep really needs sudo. install it.
No sudo in xenial image
No sudo in xenial image
fix ubuntu version. and again.
fix ubuntu version
Also build on Ubuntu 16.04
fix: no default values
I think this is it.
say yes to apt-get. all the time
No debug symbols installed
build ceres from source
We can force the dpkg installation
handle install with apt-get force
Next Ubuntu hack
fix stupid Ubuntu typo
Introduce hack because Ubuntu
rosdep really needs sudo. install it.
No sudo in xenial image
No sudo in xenial image
fix ubuntu version. and again.
fix ubuntu version
Also build on Ubuntu 16.04
Merge branch 'Test' into 'master' Update AUTHORS See merge request !89
Merge branch 'libsvm' into 'master' fix libsvm dependency in package.xml See merge request !90
fix libsvm dependency in package.xml
Update AUTHORS
Merge branch 'add_boost_dependency' into 'master' Update hypotheses_verification.cpp (wtf? I just commited this change) See merge request !88
Update hypotheses_verification.cpp
Update hypotheses_verification.cpp (wtf? I just commited this change)
Merge branch 'master' into 'master' Master See merge request !82
Merge branch 'add_boost_dependency' into 'master' add missing boost dependency See merge request !86
add missing boost dependency
Go back to use standard ubuntu trusty docker image This is easier to support in the future.
Merged branch master into master
Merged branch master into master
Merge branch 'fix_vector_type' into 'master' Fix vector type See merge request !79
Update hypotheses_verification.cpp
Update hypotheses_verification.cpp
fix vector type
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
fix vector type for new histogram interface
Update Readme.md
Merge branch 'master' into 'master' Use docker image that has those dependencies already installed See merge request !78
Use docker image that has those dependencies already installed
Merge branch 'master' into 'master' Update .gitlab-ci.yml See merge request !77
Update .gitlab-ci.yml
Merged branch master into master
Update .gitlab-ci.yml
Update .gitlab-ci.yml
Merge branch 'some_fixes' into 'master' Some fixes See merge request !56
Update .gitlab-ci.yml
Merged branch master into master
Update .gitlab-ci.yml yet again. ;-)
Update .gitlab-ci.yml
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
put some header defintions into cpp files and remove .hpp files
Update .gitlab-ci.yml
put miscellaneous functions into more specific files
Update v4r_style_guide.md
fix merge conflict
Merged branch master into master
added: only small inline functions
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Update CONTRIBUTING.md
Merged branch master into master
added: keep pull requests short
fixed typo
fixed typo
clean up .gitlab-ci.yml
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
add depdendencies description
Update .gitlab-ci.yml
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
add contributing and style_format doc files
Update package.xml test if this compiles now
Update .gitlab-ci.yml Continue on rosdep errors. Arrrrrr
Update .gitlab-ci.yml fix syntax
Update .gitlab-ci.yml specify the ROS version (needed to resolve packages from package.xml)
Update .gitlab-ci.yml We need wget as well.
Update .gitlab-ci.yml
Update .gitlab-ci.yml Seems like we need cmake after all
Add .gitlab-ci.yml First try
update color transformation and comparison
use boost::dynamic_bitset instead of vector<bool>, add camera class, put some definitions into header files + some code refactoring
fix roi when reaching boundary
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'master' into 'master' Ubuntu 16.04 compatibility See merge request !54
Merge pull request #67 from strands-project/ubuntu1604_compatibility Ubuntu1604 compatibility
Merge remote-tracking branch 'hannes/master'
add pcl time header
Merge remote-tracking branch 'v4r-master/master'
add bounding box function
seperate definitions from some header files to reduce compile time
fix some warnings
Merge pull request #66 from strands-project/sync_from_gitlab Sync from gitlab
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'new_try' into 'master' New try See merge request !53
Merge pull request #64 from strands-project/new_try [WIP] New try
add timing for pose refinement
update citation file
put pcl_opencv functions from header into implementation file
some code optimization
some changes for compiling with Ubuntu 16.04
some beauty
add script for obtaining alexNet CNN
use const
make destructors virtual for virtual classes
remove empty file
remove broken files
put test data into directory
fix existing directory warning
add docs for recognition update get_TUW script
add script for downloading 3dnet test data
add doc for RTMT
include scripts to obtain training data from TUW and 3dNet
include missing mean substraction in alexnet feature estimation
update for shape cnn classifier to work
fix wrong model assembly resolution
fix compilation errors for eval and app pieces
fix missing clear of indices when no keypoints are detected
remove voxelgriddistancetransform method call
remove default typename in createIndicesFromMask method to allow usage without c++11
add global hypotheses non-maxima surpression to rejection method
group hypotheses by segmentation cluster
add online averaging function
add hyp stuff (should have been staged earlier)
remove EDT stuff
check if all views are trained during initialization (not just if directory exist)
put boost program options directly into parameter classes, merge ghv with hypotheses verification class
make seperate table class
minor fixes for save pose into pcd
update some visualization functions in recognition
remove sift based alignment in incremental object learning class
use new segmentation class and provide combined feature, keypoint and segmentation layer
hopefully fixing Caffe optional dependency
up pcl version
fix compilation error caused by addcoordinatesystem if used with PCL < 1.7.2
add esf classifier again
fix typo in openmp call
fix some warnings
fix bug in optional dependening on caffe and rendering
change default params and do not instantiate harris and iss keypoint extractor on PCL versions < 1.7.2 (as keypoint indices is not available for these versions)
make recognition library dependency to rendering and Caffe optional
move some hpps into cpps
skip recongition rate computation if output file already exists
add nicer visualization for paper
add todo comment
add eval for rec rate over occlusion
fix crop bug in pcl opencv conversion
fix min fitness threshold scaling
flip table plane towards viewpoint and make parameter for min points accessible
make resolution as an integer in mm
add coordinate system for visualizing recognition results
fix bug in color conversion
change default parameter for svm cross validation
make smooth segmentation parameter scale with depth
avoid table plane filtering for initialization in local recognizer
add parameter options for smooth clustering
add dense SIFT option (not tested yet and only available for SIFTGPU)
add smooth clustering and linear thresholding of model fitness threshold (with visibliity)
use multi-plane segmentation for local recognizer to find heighest table plane
fix visualization for recognition rate computation when cloud sensor header is changed
temporary remove parallel call of recognizer QH6205 qhull error (qh_initqhull_start): qh_qh already defined. Call qh_save_qhull() first
fix bug in compute recognition rate
ignore multiview and esf object classifier for now
make model fitness threshold adaptive to visible ratio (TODO: write a proper function)
use bigger rendering points in model cues visualization
fix wrong sigma for AB color components
remove table plane debug visualization
rename some recognition init parameters
reset view to do not mess up visualization in evaluation recognition example
add option to just compute dominant plane (without clustering)
fix bug with multiple call to recognize if recognizer is local
add all the options for initialization
make local recognizer more modular
fix bug in knn_sift initialization
add missing iostream include in classifier.h
add opencv sift option again (NOTE: Not tested yet)
remove keypoint rounding stuff in sift
rewrite local estimator interfaces
remove redundant files, take into acccount sign ambiguity of eigen vectors for global recognizer
fix bug with missing normal computation
migrated feature estimator changes (except eigen matrix). kinda working but only for first test view it seems
add global recognizer
add ourcvfh pcl trunk version, fix view all point clouds in folder
merging svmwrapper, classifier classes, keypoint extractors... still working
merged many things from rec_optimization_isolated branch (hyp generation still working - verificaiton not)
add point cloud to eigen matrix conversion with indices
add ptr for gcg
remove old comments
add ptr for gcg
fix merge conflict
add vector sort with indices return
add some histogram functions
add cielab2rgb transformation
some code polish in graph based geometric consistency grouping
avoid some warnings
add visualization of model keypoints
fix visualization of correspondences
remove global requirement for samples to have all modules enabled it now only looks for the individual dependency of each sample and compiles just the ones which meet their dependencies
using parameter class for gcg when gcg is used... small code polish
addition to previous commit
fix color conversion compilation error in case of PointXYZ instantiation
make ghv compile for PointXYZ type instantation as well
fix error with color retrieval in verification code
optimize speed
fix bug in model assembly
remove parameter output
speed up verification evaluation by compressing scene fitting matrix
add recognition rate evaluation
make it compile for PointXYZ as well
compute visible model cloud on multiple views
merge hv_go3d into ghv (not ready yet) optimize visible model cloud computation in verification
add depth image computation in zBuffering class (remove XYZRGBA template instantition)
split code into more functions, add omp sections again, and some minor beauty
add replace moves again by checking pairwise intersection matrix
enhance pairwise intersection computation by fixing smoothing, speeding up computation and adding erosion
remove some more obsolete code
use new verification cost function and remove obsolete code pieces
add smoothing function to zbuffering (does not work properly though)
add function to remove column or row from eigen matrix
fix compiler error in change detection module
implement pairwise intersection computation in verification algorithm
add rendering function in zbuffering (explicit)
use local point color to compare color
delete obsolet member variables
delete count active hypotheses function in verificitation (as it is not used anyway)
make update function use member variables instead of having to pass them as an argument
do not use weights for outliers - just ratio of number of outliers compared to visible points
clip noise model based radius for inliers search
rename variable and do label check earlier to avoid redundant processing
fix seg fault when not using icp for pose refinement
reset camera view point in object recognizer to avoid messing up visualization
clip max viewing angle in noise model to 80 degrees to avoid huge noise terms (was 85)
use noise model for model explained points
fix ignore color even if exists check
fix wrong use of row and column counter in self zbuffering module
do incremental smooth clustering via noise model (not finished yet)
make visualize go cues a switch parameter
add a static function to query noise level for a single point
temp commit
Integration of change detection into recognition module
Annotation of changes in GT data
Change detection module added
Compilation fix: duplicated pragma
add merge for multiview
normalize optimization variables
fix multipipeline merging of hypotheses when disabled. Also skip merging of ident hypothesis
fixed self occlusion reasoning
add pose refinement fix noise model based cloud integration for just one input cloud as well as for no indices given
fixed points on plane side variable in ghv
working again
explained and unexplained points seem okay
fix merge conflict
fix merge conflict
Merge pull request #63 from taketwo/remove-x86 Remove all mentions of x86 and x86_64 in CMake scripts
Remove all mentions of x86 and x86_64 in CMake scripts
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
use object indices also for unfiltered registered model cloud and only save filtered input clouds if debug option is set
Merge branch 'master' into 'master' added quick fixed to handle some range check exceptions needs proper handling soon See merge request !51
1.3.3
1.3.2
Merge remote-tracking branch 'upstream/master'
add missing Xxf86vm lib
Merge remote-tracking branch 'remotes/upstream/recognition_update'
Contributors: Georg, Georg Halmetschlager-Funek, Johann Prankl, Markus Bajones, Markus Suchi, Martin Velas, Michael Zillich, Sergey Alexandrov, Simon Schreiberhuber, Thomas Fäulhammer

1.4.3 (2016-02-26)

1.4.2 (2016-02-26)

Merge pull request #60 from strands-project/strands some quick fixes regarding range check exceptions, need proper fix eventually
Merge branch 'master' of github.com:strands-project/v4r into strands
added quick fixed to handle some range check exceptions needs proper handling soon
Merge pull request #59 from strands-project/fix_range_error_when_using_hv_use_histogram_specification Update ghv.h
Update ghv.h
Contributors: Michael Zillich, Thomas Fäulhammer, mzillich

1.4.1 (2016-02-01)

Merge pull request #58 from strands-project/fix1 initialize counter variable
initialize counter variable
Merge pull request #57 from strands-project/remove_c+11_from_header remove c++11 construct in header file
remove c++11 construct in header file
Merge pull request #56 from strands-project/fix1 Fix1
add siftgpu as optional dependency in RTMT
copy uniform_sampling files from PCL 1.7.2 to make V4R also compile on PCL 1.8
updated RTMT noise model parameters
Merge remote-tracking branch 'v4r_root/master'
Merge branch 'dynamic_object_learning' into 'master' Dynamic object learning See merge request !50
Merge branch 'master' into 'master' Master See merge request !49
Contributors: Thomas Fäulhammer

1.4.0 (2016-01-27)

Merge pull request #55 from strands-project/new_recognition_resolved_merge_conflict New recognition resolved merge conflict
Merge branch 'new_recognition'
Merge remote-tracking branch 'strands/master'
change default values
fix noise model based cloud integration
make opencv sift instantiation conditional on siftgpu presence
integrate parse console arguments into library
Merge pull request #54 from taketwo/speed-up Speed-up info collection in NMBasedCloudIntegration
uses more parallelization
Merge remote-tracking branch 'sergey_strands/speed-up' into new_recognition
Speed-up info collection in NMBasedCloudIntegration Pre-compute the number of points and resize big_cloud_info_ only once. This achieves > 2x speed-up in Debug mode.
tmp commit to test siftgpu
some beauty
add present of model in view variable for go3d change default noise model param
parallelize add models function in go3d
some beauty
normalize all components of LAB colors in range -1 to 1
put color transform into seperate class
remove a few pointers and add parallel block refactor code for merging feature correspondences in multiview recognizer
fix conditional compilation with -DWITH_SIFTGPU=OFF
remove hough_3d as it is not used and within PCL (maybe other version though)
remove accidentally added build folder
remove template parameters FeatureT and DistT for local recognizer/estimator save descriptors as binary text file on disk
getting rid of some pointers move duplicated functions in a common file
make multipipeline recognizer code parallel
parallelize correspondence grouping
make converttoflann create its flann data internally (to make interfacing simpler)
hopefully solves range_check_error during correspondence grouping refactored some code
add missing ifdef HAVE_SIFTGPU
fix interface problem in IOL and avoid deprecated interface
Merge pull request #52 from strands-project/add_citation_license_file add citation, license and authors file
add citation, license and authors file
Merge pull request #51 from strands-project/build-fixes Build fixes
Merge remote-tracking branch 'severin/build-fixes' Conflicts: samples/examples/object_recognizer_new.cpp
Merge pull request #49 from strands-project/fix_siftgpu_problem_in_IOL Fix siftgpu problem in iol
use HAVE_SIFTGPU to check if siftgpu is available on system in object modelling module
rename dynamic object learning to incremental object learning
Added missing header 'boost/format.hpp' in a few examples
[cmake] ObjectGroundTruthAnnotator requires VTK
[cmake] Ensure v4r compiles without ceres at CMake level Note that V4R does not yet compile without ceres due to modules/reconstruction/include/v4r/reconstruction/impl/ReprojectionError.hpp requiring ceres.h
Properly guards omp.h includes so that the project compile without OpenMP support
[cmake] Cosmetic in CMakeLists
[cmake] Use pkg-config to find OpenNI2 The Debian package for libopenni2 provides a .pc but no FindOpenNI2.cmake
[cmake] FindOpenGL does not return the version
[cmake] Added support for compiling with Qt5 Note that CMake option WITH_QT needs to be enabled, and WIT_QT4 needs to be disabled.
[cmake] Enable WITH_QT by default
Merge pull request #44 from strands-project/dynamic_object_learning Dynamic object learning
make compatible to new v4r interfaces
Merge branch 'dol_rebased'
Merge branch 'master' of github.com:strands-project/v4r
fix deprecated warning
remnants from RAL paper
fixed some bugs
fix of fix
fixed bug in evaluation - don't test on same set as object was trained
write eval recognition for uncontrolled scenes to test on controlled ones added visualization of all learnt models
fix wrong parameter type
added file to test model coverage
skip patrol run for which no object has been modelled
recognition evaluation with respect to coverage for controlled runs
Merge branch 'recognition_dev' into dol_rebased Conflicts: modules/recognition/include/v4r/recognition/impl/local_recognizer.hpp
taken view file is now correct
forgot to undo temporary change
added evaluation tool for recognition performance measure of partial model coverage
add FindX11 and FindXRandR
fixed error when training views do not begin with 0
recognition evaluation for partial model and for offline data more or less ready
added first evaluation tool to compute recognition performance with respect to percentage of visible model
added visualize_masked_pcd again
Merge branch 'recognition_dev' into dol_rebased
Merge branch 'dol_rebased' of rgit.acin.tuwien.ac.at:t-faeulhammer/v4r into dol_rebased Conflicts: samples/examples/dynamic_object_learning.cpp samples/icra16/eval_dol_gt.cpp samples/icra16/eval_dynamic_object_learning_with_mask_pertubation.cpp
add eval use boost program options
adapt code to make rebase compile
rebase commit
added noise level evaluation for initial mask ICRA16
added eval for inital mask evaluation added for icra16 singleview
fixed sv eval when test_dir is not present
fixed bug in icra sv eval, when csv file has only 2 columns
eval almost ready
added icra vis
seperate post-processing and save to disk in object learning
fixed wrong output file if name of mask is mask.txt only
removed overhead computation when sift based camera pose estimation is disabled
fixed ground truth labelling
fixed color in add text
just addded a const
removing nan points in initial mask - otherwise seg fault when after erosion not enough points
included plane merge moved logical stuff to common module added plane visualization added plane properties
added function to write images to disk for intermediate steps
make ratio parameter accessible from outside for occluded and object supported points
sort files before evaluation and output debug info
added some V4R_EXPORTS in registration module removed redundant fast_icp in common module added app for 3D reconstruction based on SIFT and MST fixed CERES version conflict fixed some dependency issues
up
add Willow Dataset definition for save_pose_into_pcd sample
set sensor pose to identity in eval to show right visiualization
parameters can now also be set in constructor initial eval code now in samples (should be moved somewhere else later on)
moved mask<->indices conversion function into v4r common module
added object_modelling again
adapt code to make rebase compile
rebase commit
added noise level evaluation for initial mask ICRA16
added eval for inital mask evaluation added for icra16 singleview
fixed sv eval when test_dir is not present
fixed bug in icra sv eval, when csv file has only 2 columns
eval almost ready
added icra vis
seperate post-processing and save to disk in object learning
fixed wrong output file if name of mask is mask.txt only
removed overhead computation when sift based camera pose estimation is disabled
fixed ground truth labelling
fixed color in add text
just addded a const
removing nan points in initial mask - otherwise seg fault when after erosion not enough points
included plane merge moved logical stuff to common module added plane visualization added plane properties
added function to write images to disk for intermediate steps
make ratio parameter accessible from outside for occluded and object supported points
sort files before evaluation and output debug info
added some V4R_EXPORTS in registration module removed redundant fast_icp in common module added app for 3D reconstruction based on SIFT and MST fixed CERES version conflict fixed some dependency issues
up
add Willow Dataset definition for save_pose_into_pcd sample
set sensor pose to identity in eval to show right visiualization
parameters can now also be set in constructor initial eval code now in samples (should be moved somewhere else later on)
moved mask<->indices conversion function into v4r common module
added object_modelling again
Contributors: Sergey Alexandrov, Séverin Lemaignan, Thomas Fäulhammer

1.3.1 (2016-01-13)

Merge pull request #43 from strands-project/fix_classifier Fix classifier
fix global classifier error when reading from new model database file structure
build utility tools by default
Merge remote-tracking branch 'simon/master'
Merge pull request #42 from strands-project/remove_glfw3_dependency Remove glfw3 dependency
remove some output messages
remove glfw3 dependency and use X* libraries only fixed some deprecated interfaces warnings added some build /run dependency for openni
Added code for a proper destructor
cleaned up some code
Merge remote-tracking branch 'simon/master'
Merge remote-tracking branch 'simon/change_glfw_to_old'
removed the need for glfw and changed everything to work with only x11 dependencies
Merge remote-tracking branch 'origin/master'
Merge remote-tracking branch 'v4r_root/master'
Merge remote-tracking branch 'strands/master'
merged
use openni instead of openni2
Merge remote-tracking branch 'v4r_root/recognition_dev'
Merge branch 'recognition_update' into 'master' Recognition update See merge request !2
Merge branch 'recognition_update' into 'master' Recognition update See merge request !45
Contributors: Simon Schreiberhuber, Thomas Fäulhammer

1.3.0 (2016-01-08)

1.2.0 (2016-01-08)

Merge pull request #40 from strands-project/recognition_dev Recognition dev
use openni instead of openni2
rename object tracker
updated object tracker and RTMT saves tracking model correctly
seperated normal computation methods into new file using using namespace v4r in samples for (i) brevity, (ii) conformity with ROS wrappers changed some deprecated interfaces split header files into impl (.hpp) files mainly to avoid c++11 neccessity when including with ROS
temporary backup commit
noise model based cloud integration update (also moved to registration module) uses properties (1) lateral noise, (2) axial noise, (3) distance in px to depth discontinuity
backup commit
first try for new noise modelling
Merge branch 'fix_KeypointSlamRGBD' into recognition_dev
fixed problem when training views do not start with cloud_000000.pcd TODO: re-initialize problem still exists (if training database is altered, flann matrix will be wrong - have to remove *_flann.idx manually right now) fixed trigraph warnings
use absolute value when checking reprojected poitns in ground truth annotation added parameters for noise model based integration demo program
since image2 only takes integer values, we do not need to interpolate (checked by Hannes)
add zero padding in interpolationfunction to avoid assertion error in Debug mode add fix from Hannes
fix seg fault when dist_coeffs is 2x4 matrix instead of 1x8
small fix (avoid ourcvfh)
rewrite noise model based integration so that it uses really equation from Nguyen et al paper.
tmp commit
tmp commit
replaced a few integer and long unsigned integer by size_t to hopefully make it working on 32bit machines added visualization functions for hypotheses verification
some more changes in pcl2opencv
change pcl2opencv interfaces
added tools again
Merge remote-tracking branch 'strands/master' into add_v4r_exports
added a few more V4R_EXPORTS (visibility attribute) for classes added cmake_minimum_required version (cmake 2.8.8 apparently can not handle url hash tags)
add v4r_export for tomita
Contributors: Thomas Fäulhammer

1.1.1 (2015-11-23)

Merge pull request #37 from strands-project/add_glm_run_dependency add glm also as run dependency
add glm also as run dependency
Contributors: Thomas Fäulhammer

1.1.0 (2015-11-20)

Merge pull request #35 from strands-project/recognition_update Recognition update
Merge remote-tracking branch 'v4r_root/recognition_update' into recognition_update1
Merge branch 'fix_glfw3' into 'recognition_update' Fix glfw3 and undefined references to X* See merge request !47
add multiple X11 linker libs
Fix variable names in examples CMakeLists
Merge remote-tracking branch 'sergey/fix-glfw3' into recognition_update1
Export 3rdparty include directories into config file
undo insert HAVE_V4R_RENDERING
add some x*libraries in package.xml to hopefully solve undefined references
Merge remote-tracking branch 'sergey/fix-glfw3' into recognition_update1
added description for go3d parameter
Properly add GLFW3 third-party library
Merge branch 'fixes-for-recognition-update' into 'recognition_update' Fixes for recognition update This fixes a few compilation problems in the current recognition update branch. See merge request !46
Fix "invalid suffix 'd' on floating constant" error
Add missing dependency (rendering depends on GLM)
added glog as dependency (otherwise linking potentially fails)
updated parameters for sv recognition
added conversion function from point cloud to fixed sized image removed unused parameters in global estimator changed namespace of pclopencv to v4r
computing histogram size by sizeof to make esf estimator compatible with PCL 1.7.1
remove template parameter Feature from global classifier and make it a vector instead added esf object classifier again
tmp commit
Merge remote-tracking branch 'simon/recognition_update' into recognition_update1 Conflicts: modules/rendering/src/depthmapRenderer.cpp
tmp commit (conditional saving of pcd as xyz or xyzrgb) before merging simons update
Cleaned up the code and sorted out some culprits.
fixed datatype for colormap fixed some warnings added program options for radius, subdivision, camera intrinsics,...
added glGetError queries.
added rendering + example added glew, glfw find package
updated some more parameter descriptions
renamed occlusion reasoning into zbuffering and removed second layer namespace seperated classes into seperate files renamed boost graph extenstion into multi-view representation fixed shot recognizer (removed indices), parameters are now written to file
added GLOG dependency using boost program option for object recognizer examples and Ground-truth annotator
use integer key for model assembly (instead of float) - resolution for it is now a parameter temporary included visualization for pose refinement
parameters are now double (instead of float) to be easily accessible from outside via ros getparam default parameters change updated ground truth annotator and evaluations for recognizer to new framework
added clear multiview data
Properly export template instantiations in EDT
Fix METSlib third-party library
removed visualization reminiscent in single-view recognizer
fixed wrong index computation in 3D occupancy grid removed siftgpu library from necessary dependency in reconstruction app
fixed wrong angle difference calculation when clustering object hypotheses [TODO: make parameter accesible from outside]
(hopefully) fixes crash when no valid recognition model is found merging close hypotheses is now possible (generate less hypotheses, refines pose by these larger correspondence set)
using mask instead of indices in go3d addModels function increased default occlusion threshold can be compiled with clang again
fixed multiplane segmentation in unorganized point clouds (TODO: downsample cloud) replaced USE_SIFT_GPU definitions with HAVE_SIFTGPU v4r_config.h now presents status of HAVE_SIFTGPU
added pcl version of ClusterNormalsToPlane (works for unorganized point clouds now) TODO: fix multiplane segmentation method
install metslib header files fixed go3d createVoxelGridAndDistanceTransforms is now called inside generate (for registered views source) TODO: Check influence of resolution paramter! added some description and licenses
temporary commit with GO3D visualization
fixed wrong transformation of keypoints when using -transfer_feature_matches 1
added mising tracking dependency
recognizer: - added license - removed unused variables - moved internally used public methods to protected
go3d implemented but results not satisfying (parameter not good?)
pruningGrap in Multiview Object Recognizer is working [TODO: Finish Go3D Hypothesis Verification Integration]
failed try of point cloud rendering with vtk
when using compute_mst, it crashes after using pruneGraph
absolute pose computation seems to work
absolute pose computation seems to work
added merging of feat correspondences
tmp commit
temporary commit (single-view recognizer correspondence grouping seems broken)
adding parameter classes remove redundant variables getting rid of singleview_object_recognizer class local estimator uses normal estimator from v4r common now Reimplementation of multiview recognizer just started (NOT WORKING / COMPILING)
single view object recognizer almost ready
tmp commit getting rid of redundnant single_view object recognizer class
correspondences in recognizer are now stored as indexes to original cloud this should reduce memory requirement New parameter class for Hypotheses Verification methods (different results to before - TODO: double check default parameters!)
only training dir parameter neccessary any more improved code readability
temporary commit (signatures not initialized) otherwise it seems to work
overall update of training procedure
recognizer structure sift parameter was named inconsistently fixed some warnings
this includes the changes from gitlab v4r version made by @alexandrox88 - fixes assimp in tomgine - remove ipp adds object tracking fixes a few warnings
SOMETHING SEEMS TO BE WRONG WITH THE TRANSFORMS namespace update polishing multiview recognizer add libsvm as system dependency
merged remove_tomgine
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
Merge branch 'remove-ipp' into 'master' Remove all mentions of IPP (Intel Performance Primitives) Remove all mentions of IPP (Intel Performance Primitives). This remained from OpenCV scripts. See merge request !43
Remove all mentions of IPP (Intel Performance Primitives)
Merge branch 'fix-tomgine-assimp' into 'master' Fix Assimp dependency This merge request fixes missing Assimp include in Tomgine and updates the CMake script for finding the package. See merge request !42
Update Assimp finder script
Add missing AssImp include in tomgine
Fix a few warnings in tomgine
Merge branch 'master' into 'master' Master created a tracking module and added the monocular object tracker from RTMT See merge request !41
Merge branch 'find-system-libsvm' into 'master' Add CMake commands to detect system installation of LibSVM The possibility to build LibSVM from source is preserved, but has to be enabled by setting BUILD_LIBSVM option (which is now off by default). See merge request !40
added monocular camera pose tracker (lk/ keypoint based) from RTMT
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
test
test
mv test
just a test file
Contributors: Johann Prankl, Markus Bajones, Sergey Alexandrov, Thomas Fäulhammer, simon.schreiberhuber@gmx.net

1.0.11 (2015-10-14)

Merge pull request #34 from strands-project/remove_tomgine temporary remove Tomgine and everything related to it (i.e. object cl…
also comment computeCentroid in single-view object recognizer
comment computeCentroid to silence error
temporary remove Tomgine and everything related to it (i.e. object classification)
Contributors: Thomas Fäulhammer

1.0.10 (2015-09-21)

Merge pull request #31 from strands-project/namespace_update Namespace update
namespace update polishing multiview recognizer add libsvm as system dependency
Merge remote-tracking branch 'sergey/find-system-libsvm' into namespace_update
Add CMake commands to detect system installation of LibSVM The possibility to build LibSVM from source is preserved, but has to be enabled by setting BUILD_LIBSVM option (which is now off by default).
rename multiview_object_recognizer silence unused variable warning removed unneccessary point cloud copy normal method now a parameter
Merge branch 'master' into 'master' Master See merge request !39
Merge branch 'master' into 'master' Master See merge request !38
Contributors: Sergey Alexandrov, Thomas Fäulhammer

1.0.9 (2015-09-17)

Merge branch 'master' of github.com:strands-project/v4r
fix Bloom issue with umlauts
Merge remote-tracking branch 'strands/master'
Contributors: Thomas Fäulhammer

1.0.8 (2015-09-17)

Merge pull request #28 from strands-project/remove_c++11_flags_and_common_namespace remove C++11 flags
remove C++11 flags remove common namespace remove duplicated files divide samples in examples, evaluation and utility tools (enable examples by default in cmake) add Qt Cache files in .gitignore list
Contributors: Thomas Fäulhammer

1.0.7 (2015-09-16)

Merge pull request #27 from strands-project/new_samples_structure New samples structure
Merge pull request #26 from strands-project/add-tomgine Add tomgine
new samples structure
divide samples into examples, tools and evals
adds ESF classifier using new point cloud rendering based on TomGine (camera pose is not extracted right now)
Merge pull request #24 from strands-project/sift_gpu_solution Sift gpu solution
added initial segmentation example
updated usage output
added tomgine
added Random Forest and SVM
Merge remote-tracking branch 'sergey/add-libsvm' into add-libsvm
added RandomForest fixed some warnings
Add libsvm 3rd-party library
Merge branch 'master' into 'master' Master See merge request !37
reverted sv recognizer header file because otherwise cg pointer cast caused seg fault fixed some warnings
make SIFT_GPU optional by setting BUILD_SIFTGPU in cmake
added segmentation dependency to samples
added binary vector increment changed parameter name to avoid confusion in range image computation
merged
Merge branch 'master' into 'master' Master this hopefully includes all the changes from LaMoR Summer School + fixes for the Recognizer See merge request !36
Contributors: Sergey Alexandrov, Thomas Fäulhammer

1.0.6 (2015-09-07)

Merge pull request #23 from strands-project/mergeLAMOR Merge lamor
merged lamor STRANDS
Merge branch 'master' of github.com:strands-project/v4r into mergeLAMOR
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r into mergeLAMOR
added default param for printParams in MV recognizer other minor changes
Update Readme.md
hopefully fixes bug in ourcvfh with different pcl versions view_all_point_clouds_in_folder can now also save images to disk
Merge branch 'master' into 'master' Master See merge request !35
catch SIFT FLANN exception when updating model database
flann idx now configurable
Merge branch 'master' into 'master' Master See merge request !34
Merge branch 'master' into 'master' Master See merge request !33
Contributors: Marc Hanheide, Thomas Fäulhammer

1.0.5 (2015-08-30)

1.0.4 (2015-08-29)

Merge pull request #22 from strands-project/marc-hanheide-patch-1 disable C++11
disable C++11 see https://github.com/strands-project/v4r_ros_wrappers/commit/0f008ac162ef2319d5685054023ce2c6f0c8db55
disable C++11 see https://github.com/strands-project/v4r_ros_wrappers/commit/0f008ac162ef2319d5685054023ce2c6f0c8db55
Contributors: Marc Hanheide

1.0.3 (2015-08-29)

Merge pull request #21 from strands-project/added_install_commands added install targets for apps
added install targets for apps
Contributors: Marc Hanheide

1.0.2 (2015-08-29)

Merge pull request #20 from strands-project/marc-hanheide-patch-1 don't include FREAK headers
don't include FRAK headers as this seems to fail in non-free opencv... see strands-project/v4r_ros_wrappers#3
Contributors: Marc Hanheide, Michael Zillich

1.0.1 (2015-08-28)

fixed some compiler warnings fixed out of range bug in GHV RGB2CIELAB when RGB color is white (255,255,255) fixed typo in parameter for eval sv
removed comments in sv recognizer, save parameter file in sv recognizer eval
removed linemod and debug build for recognition
fixed bug in sv_recognizer
added EDT include path
added ground truth annotator as app removed unused files in recognition
added sv recognition sample fixed missing chop_z behaviour in single view recognizer
added sample eval for single view object recognizer
updated ReadMe
added libglm-dev as build dependency
Merge branch 'add-glm-dependency' into 'master' Add GLM dependency See merge request !32
Add GLM dependency
Merge branch 'master' into 'master' Master See merge request !31
added cmake files for openni2
Merge branch 'master' into 'master' Master See merge request !30
Merge branch 'fix-u-r' into 'master' Fix undefined reference errors (with Octree and 1.7.1) See merge request !29
added qt-opengl-dev as dependency
added openni in package.xml
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
linked openni libraries to RTMT added octree_impl to hopefully solve pcl conflicts with versions <= 1.7.1
Hopefully fix undefined reference errors (with Octree)
Merge branch 'add-template-keyword' into 'master' Add missing 'template' keyword (causes clang compilation error) See merge request !28
Merge branch 'master' of rgit.acin.tuwien.ac.at:root/v4r
added RTMT GL libraries again
Add missing 'template' keyword (causes clang compilation error)
added binary operations to common changed dist calculation for planes available from outside
fixed QT4 CMake file
Merge branch 'master' into 'master' fixed QT4 CMake file See merge request !27
Merge branch 'master' into 'master' added RTMT See merge request !26
added RTMT
Merge branch 'master' into 'master' Master See merge request !25
added possibility to crop image when converting PCD to image createDirIfNotExists should now create all directories recursively added initial version for pcl segmentation (used in STRANDS in Year1) - not finished
make parameters double (instead of float) to make it directly accessible via ros getparam function
Merge branch 'master' into 'master' Master See merge request !24
fixed error with Willow Poses removed object modelling dependency which is not yet present
added const specifier for get function
Merge branch 'master' into 'master' Master See merge request !23
Merge branch 'master' of rgit.acin.tuwien.ac.at:t-faeulhammer/v4r Conflicts: samples/cpp/save_pose_into_pcd.cpp
added some V4R_EXPORTS in registration module removed redundant fast_icp in common module added app for 3D reconstruction based on SIFT and MST fixed CERES version conflict fixed some dependency issues
fix of last push
fix of last push
added definitions for willow_dataset in save_pose_into_pcd sample
added mask<->indices converter function ground truth annotator now also outputs mask for object in first frame
added initial version for ground truth labelling tool
del
added samples folder
Merge branch 'dynamic_object_learning' of rgit.acin.tuwien.ac.at:t-faeulhammer/v4r into dynamic_object_learning
fixed some ns
fixes some namespace issues
added object learning again
fixed pcl version conflict with vector of eigen
Merge branch 'master' of rgit.acin.tuwien.ac.at:t-faeulhammer/v4r
fixed vector conflict with different PCL versions
fixed some ns
Merge branch 'master' into dynamic_object_learning
changed ns
fixed wrong macro names for detect CUDA cmake
Merge branch 'dynamic_object_learning' of http://rgit.acin.tuwien.ac.at/t-faeulhammer/v4r into dynamic_object_learning
Merge branch 'dynamic_object_learning' of rgit.acin.tuwien.ac.at:t-faeulhammer/v4r into dynamic_object_learning
added object learning again
fixes some namespace issues
Merge branch 'dynamic_object_learning' of rgit.acin.tuwien.ac.at:t-faeulhammer/v4r into dynamic_object_learning
added object learning again
fixed wrong cmake macro name
added object learning again
del
Merge branch 'master' into 'master' del See merge request !22
Merge branch 'master' into 'master' Master See merge request !21
remnoved second layer namespace "rec_3d_framework" added some V4R_EXPORTS changed some include paths removed redundant faat_3d_rec_framework.h file
Merge branch 'dependency-propagation' of rgit.acin.tuwien.ac.at:alexandrov88/v4r into dependency_propagation
Print OpenCV and Ceres statuses
Update find Ceres to export variables
Implement dependency propagation
Split filesystem_utils into parts
Remove duplicate find eigen call
Properly set variables in FindEDT
Properly set variables in FindOpenCV
Properly set variables in FindEigen
SiftGPU fixup
Boost fixup
Change SIFTGPU_INCLUDE_DIR -> SIFTGPU_INCLUDE_DIRS
Update io module
Find Boost globally
Merge branch 'master' into 'master' Master See merge request !20
Merge branch 'master' into dynamic_object_learning
added camera tracker - uff, so many changes!
updated recognition cmakefile to have correct link to opencv fixed some shadow warnings
fixed some warning and added V4R_EXPORTS added openmp in cmake
Merge branch 'master' into 'master' fixed some warning and added V4R_EXPORTS added openmp in cmake See merge request !19
Merge branch 'fix-edt' into 'master' Build EDT library with -fPIC option See merge request !18
Build EDT library with -fPIC option
fixed some warnings changed default parameter value of sor
Merge branch 'master' into 'master' Master See merge request !17
Merge branch 'master' into dynamic_object_learning
added object_modelling cmakelists.txt
added OpenCV as cmake dependency added some V4R_EXPORTS re-inserted computeOccludedPoints (why was this not merged?? Check other files!) added OpenMP cmake c/cxx flags
fixed warnings of shadowed variables using new v4r namespaces
Merge branch 'master' into dynamic_object_learning Conflicts: modules/object_modelling/include/v4r/object_modelling/do_learning.h modules/object_modelling/include/v4r/object_modelling/model_view.h modules/object_modelling/src/do_learning.cpp modules/object_modelling/src/visualization.cpp
updated EDT include path
Merge remote-tracking branch 'sergey/cmake_updates'
Create core module, moved macros.h and version.h here
All modules now explicitly depend on PCL
Fix EDT
added missing segmentation dependency
Merge branch 'master' into 'master' added missing segmentation dependency See merge request !16
adapted to new cmake system
Merge branch 'master' into 'master' Master See merge request !15
Merge pull request #19 from strands-project/new_cmake New cmake
Merge branch 'master' into dynamic_object_learning Conflicts: modules/CMakeLists.txt
Fix 3rd party header handling for the case of no-install usage of V4R
Merge branch 'new_cmake' into 'master' New cmake See merge request !14
changed required PCL version to less restrictive 1.7. Otherwise, there is a conflict on Jenkins because it only provides package for 1.7.1
hide recognition module for the time being
added package.xml again - Jenkins needs it to build the library added sergey to maintainer list
Merge remote-tracking branch 'sergey/master' into new_cmake Conflicts: modules/recognition/CMakeLists.txt modules/registration/CMakeLists.txt
Fix V4RConfig.cmake for use without installation
Merge branch 'master' into dynamic_object_learning
fixed some warnings with redundant typenames and wrong derived signature (& missing) in Recognition fixed missing EDT dependency in Registration
Merge branch 'master' into 'master' Master See merge request !13
updated supervoxel clustering method added some function docs optional parameter for pairwise transform refinement
filtering smooth clusters works -- without visualization
smooth clusters work now --- with visualization for debug
Miscellaneous should not depend on keypoints
Revert "(Temporarily) move miscellaneous to keypoints because it depends on them" This reverts commit 8b4bf90048750c95bae136b9b65dbb890c8c900e.
Add V4R_EXPORTS here and there
Merge branch 'master' into dynamic_object_learning
pcl::copyPointCloud now also accepts binary obj mask
beautify code - moved from indices to mask added parameter filter_planes_only (not working for value false yet)
(Temporarily) move miscellaneous to keypoints because it depends on them
Solve undefined reference problem
Export 3rdparty include directories
Remove compatibility stuff
Finalize SiftGPU support
table filtering working now as expected... removed some unnecessary includes
temporary commit for visualizing table planes supported by object mask
Another fix for SiftGPU
Merge branch 'revert_merge_request' into 'master' Revert "Merge branch 'dynamic_object_learning' into 'master'" This reverts commit 87d034a1a8c8763657ca59ff08f9ec95a5d1c4be, reversing changes made to d183d5143b68e70de0e678a3d0659fae2a85a731. See merge request !12
Revert "Merge branch 'dynamic_object_learning' into 'master'" This reverts commit 87d034a1a8c8763657ca59ff08f9ec95a5d1c4be, reversing changes made to d183d5143b68e70de0e678a3d0659fae2a85a731.
Trying to add SiftGPU
Fix EDT
Remove SiftGPU sources
Fix EDT third-party dependency
Merge branch 'master' into dynamic_object_learning
fixed some warnings added occlusion reasoning for two clouds (optional tf) which return occlusion mask
Merge branch 'dynamic_object_learning' into 'master' Dynamic object learning See merge request !11
added parameter for statistical outlier removal (mean=50 and stddevmul=1 didn't work well on asus_box) fixed bug in CreateMaskFromVecIndices there seems to be still a problem in occlusion reasoning
Add new build system, migrate common and segmentation modules
Get rid of legacy build system stuff
Merge branch 'master' into dynamic_object_learning
Merge branch 'master' into 'master' fixed warning of unused variable in SLICO fixed visualization issue when called multiple times See merge request !10
fixed warning of unused variable in SLICO fixed visualization issue when called multiple times
updated region growing such that it does not use points already neglected by plane extractor fixed visualizition issue when calling the visualization service more than once
Merge branch 'master' into 'master' Master See merge request !9
Merge branch 'master' into dynamic_object_learning
added ceres version check updated McLMIcp.cpp to use new fixes from aitor
include devil dependency
changed to right rosdep key for glew
Merge branch 'master' into dynamic_object_learning
added some dependencies
Merge branch 'master' into 'master' Master See merge request !8
Merge branch 'master' into dynamic_object_learning
removed aitor from maintainer list
Merge remote-tracking branch 'strands/package_xml'
Merge branch 'master' into dynamic_object_learning
added parameter class for noise model based integration changed Eigen::Vector4f vector for correct allocation
indices are now stored in a more generic way visualization now also includes noise model added Statistical Outlier Removal for initial indices added logical operator for binary masks TODO: visualization does only work for first service call
added opencv dependency
fixed dependencies to the correct rosdep keys
added a first package.xml
MST is now a parameter plane indices are stored as a vector of a vector now - (otherwise high cost occured in callgrind profiler) updated clear function
Merge branch 'master' into dynamic_object_learning
createDirIfNotExist function is now in common
fixed problem with nan values (recursive absolute pose computation based on spanning tree implementation was not correct)
minimum spanning tree is working now... there are nan values transferred to nearest neighbor search -> still needs to be fixed!
bug fix - should be back to STRANDS review demo state
Merge branch 'master' into 'master' Master See merge request !7
Merge branch 'master' into dynamic_object_learning
fixed some linking problems... fixed bug in setCloudPose (last element was not set to 1) made code clang compatible...
tmp commit
Merge branch 'master' into dynamic_object_learning
fixed linking error, updated some namespaces
tmp commit
Merge branch 'master' into dynamic_object_learning
changed some recognition files to use new filesystem namespace
tmp commit
Merge branch 'master' into 'master' Master See merge request !6
temporary commit of dynamic object learning. not compiling yet!
deleted remaining temp(~) files
added keypoint files needed for object learning
added clustertonormals from keypointTools
add initial version of keypoints
Merge branch 'master' into 'master' Master See merge request !5
some fixes to merge to master
Merge remote-tracking branch 'v4r_root/master' Conflicts: 3rdparty/metslib/CMakeLists.txt CMakeLists.txt cmake/v4rutils.cmake cmake/v4rutils.cmake~ modules/common/CMakeLists.txt modules/common/include/v4r/common/noise_model_based_cloud_integration.h modules/common/include/v4r/common/noise_models.h modules/common/src/noise_model_based_cloud_integration.cpp modules/common/src/noise_models.cpp modules/recognition/include/v4r/recognition/boost_graph_extension.h modules/recognition/include/v4r/recognition/ghv.h modules/recognition/include/v4r/recognition/multiview_object_recognizer_service.h modules/recognition/src/boost_graph_extension.cpp modules/recognition/src/boost_graph_visualization_extension.cpp modules/recognition/src/multiview_object_recognizer_service.cpp modules/segmentation/CMakeLists.txt
remove ~
.
.
tmp commit
Merge branch 'master' into 'master' Added multiview recognizer. renamed some namespaces. See merge request !4
Added multiview recognizer. renamed some namespaces.
Merge branch 'master' into 'master' Master See merge request !3
Fixed merge conflict
Initial commit. For some reason if segmentation app is compiled - there is a linking problem with pcl. Namespaces are a mess!
initial commit
upd
Merge branch 'master' into 'master' update readme See merge request !1
update readme
Add new file
Init commit
Contributors: Marc Hanheide, Markus Bajones, Sergey Alexandrov, Thomas Faeulhammer, Thomas Fäulhammer

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

CHANGELOG.rst

CHANGELOG.rst

Changelog for package v4r

1.4.9 (2017-07-28)

1.4.8 (2017-04-11)

1.4.7 (2017-02-23)

1.4.6 (2017-02-14)

1.4.5 (2017-02-13)

1.4.4 (2017-02-13)

1.4.3 (2016-02-26)

1.4.2 (2016-02-26)

1.4.1 (2016-02-01)

1.4.0 (2016-01-27)

1.3.1 (2016-01-13)

1.3.0 (2016-01-08)

1.2.0 (2016-01-08)

1.1.1 (2015-11-23)

1.1.0 (2015-11-20)

1.0.11 (2015-10-14)

1.0.10 (2015-09-21)

1.0.9 (2015-09-17)

1.0.8 (2015-09-17)

1.0.7 (2015-09-16)

1.0.6 (2015-09-07)

1.0.5 (2015-08-30)

1.0.4 (2015-08-29)

1.0.3 (2015-08-29)

1.0.2 (2015-08-29)

1.0.1 (2015-08-28)

Files

CHANGELOG.rst

Latest commit

History

CHANGELOG.rst

File metadata and controls

Changelog for package v4r

1.4.9 (2017-07-28)

1.4.8 (2017-04-11)

1.4.7 (2017-02-23)

1.4.6 (2017-02-14)

1.4.5 (2017-02-13)

1.4.4 (2017-02-13)

1.4.3 (2016-02-26)

1.4.2 (2016-02-26)

1.4.1 (2016-02-01)

1.4.0 (2016-01-27)

1.3.1 (2016-01-13)

1.3.0 (2016-01-08)

1.2.0 (2016-01-08)

1.1.1 (2015-11-23)

1.1.0 (2015-11-20)

1.0.11 (2015-10-14)

1.0.10 (2015-09-21)

1.0.9 (2015-09-17)

1.0.8 (2015-09-17)

1.0.7 (2015-09-16)

1.0.6 (2015-09-07)

1.0.5 (2015-08-30)

1.0.4 (2015-08-29)

1.0.3 (2015-08-29)

1.0.2 (2015-08-29)

1.0.1 (2015-08-28)