-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathindex.html
448 lines (424 loc) · 23.6 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
<!DOCTYPE html>
<html lang="""en"></html>
<head>
<meta charset="""UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>OmniCount</title>
<link href="https://fonts.googleapis.com/css?family=Google+Sans|Noto+Sans|Castoro"
rel="stylesheet">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.0.0-beta3/css/all.min.css">
<style>
body {
font-family: 'Google Sans', sans-serif;
margin: 0;
padding: 0;
box-sizing: border-box;
background-color: #ffffff;
color: #333;
}
.container, .content {
max-width: 1200px;
margin: auto;
padding: 20px;
}
header {
text-align: center;
padding: 20px 0;
background-color: #ffffff;
box-shadow: 0 2px 4px rgba(255, 255, 255, 0.1);
}
.authors {
display: flex;
justify-content: center;
align-items: center;
flex-wrap: wrap;
background-color: #ffffff;
padding-bottom: 20px;
}
.authors a {
color: #007bff;
text-decoration: none;
margin: 0 10px;
}
.authors a:hover {
text-decoration: underline;
}
.teaser, .leaderboards, .abstract, .viz , .method, .benchmark, .bibtex, .license {
background-color: #ffffff;
padding: 20px;
border-radius: 8px;
box-shadow: 0 2px 4px rgba(130, 172, 240, 0.591);
margin-bottom: 20px;
}
.badges img {
margin-right: 5px;
}
a {
color: #007bff;
text-decoration: none;
}
/* a:hover {
text-decoration: underline;
} */
img.full-width {
max-width: 100%;
height: auto;
border-radius: 8px;
margin-top: 20px;
}
.comparison-blocks {
display: flex;
flex-wrap: wrap;
justify-content: space-around;
gap: 20px;
}
.image-block {
display: flex;
flex-direction: column;
align-items: center;
width: 250px;
}
.image-compare-container {
position: relative;
width: 100%;
height: 200px;
overflow: hidden;
}
.image-compare-container img, .overlay {
position: absolute;
width: 100%;
height: 100%;
}
.overlay {
clip-path: inset(0 50% 0 0);
}
.slider {
position: absolute;
top: 0;
bottom: 0;
width: 2px;
background: #ffffff;
cursor: ew-resize;
left: 50%;
}
.image-caption {
text-align: center;
font-weight: bold;
background-color: rgba(255, 255, 255, 0.8);
padding: 5px;
border-radius: 4px;
margin-top: 5px;
}
pre {
background-color: #007bff85;
border: 1px solid #999;
border-radius: 5px;
padding: 10px;
white-space: pre-wrap;
width: 80%;
text-align: left;
}
img {
max-width: 100%;
height: auto;
border-radius: 8px;
margin-top: 20px;
}
.icon-width {
display: block;
width: 10%; /* Adjust this value to resize the image */
height: auto; /* This is to maintain the aspect ratio of the image */
margin: auto; /* This will center the image */
}
.affiliations {
text-align: center;
}
.affiliations span {
margin-right: 10px;
}
.affiliations a {
color: black;
}
.accepted h2 {
color: #007bff;
font-size: 1.5em;
margin-top: 10px;
}
.accepted h2 a {
color: #007bff;
text-decoration: none;
}
.accepted h2 a:hover {
text-decoration: underline;
}
.container {
display: flex;
flex-direction: column;
align-items: center;
}
.section {
width: 80%;
margin: 5px 0;
}
.main-container {
padding: 0px;
margin: 20px;
box-shadow: 0 0 0 2px #007bff, 0 0 0 4px #0056b3, 0 0 0 6px #007bff;
}
</style>
</head>
<body>
<div class="main-container">
<header>
<h1><b>OmniCount: Multi-label Object Counting<br>with Semantic-Geometric Priors</b></h1>
<div class="authors">
<span><a href="https://mondalanindya.github.io">Anindya Mondal*<sup>1</sup></a></span>
<span><a href="https://sauradip.github.io/">Sauradip Nag*<sup>2</sup></a></span>
<span><a href="https://surrey-uplab.github.io/">Xiatian Zhu<sup>1</sup></a></span>
<span><a href="https://sites.google.com/site/2adutta/">Anjan Dutta<sup>1</sup></a></span>
</div>
<div class="affiliations">
<span><a href="https://www.surrey.ac.uk/"><sup>1</sup>University of Surrey</a></span>
<span><a href="https://www.sfu.ca/"><sup>2</sup>Simon Fraser University</a></span>
</div>
<div><img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/uni_icons.png" style="width: 350pt; height: auto;"></div>
<div class="columns is-centered"></div>
<div class="accepted">
<h2><a href="https://aaai.org/Conferences/AAAI-25/" target="_blank">Accepted to AAAI 2025</a></h2>
</div>
<div style="text-align: center;">
<a href="https://huggingface.co/datasets/anindyamondal/Omnicount-191" class="lnk">
<i class="fa fa-database" style="font-size: 22px;"></i>
</a> |
<a href="https://arxiv.org/pdf/2403.05435.pdf" class="lnk">
<i class="fas fa-file-pdf" style="font-size: 24px;"></i>
</a> |
<a href="https://github.com/mondalanindya/OmniCount/tree/main/code" class="lnk">
<i class="fab fa-github" style="font-size: 24px;"></i>
</a>
</div>
</header>
<div class="container">
<div class="section">
<p style="text-align: center; font-weight: bold;">TL;DR: OmniCount introduces a novel framework and the OmniCount-191 dataset for accurately counting multiple object categories in a single pass using pre-trained semantic and geometric estimation models.</p>
</div>
<div class="section">
<h2 style="text-align: center;">Object Counting Paradigms</h2>
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/omnicount_teaser.png" alt="OmniCount Teaser" class="full-width">
<p style="text-align: center;">(a) Typical single-label object counting models supports open-vocabulary counting but processes only a single category one time.
(b) Existing multi-label object counting models are training based (i.e, not open-vocabulary) approaches and also fail to count non-atomic objects, e.g. grapes.
(c) We advocate more efficient and convenient multi-label counting that is training-free, open-vocabulary and supports counting all the target categories in a single pass.</p>
</div>
<div class="section">
<h2 style="text-align: center;">Abstract</h2>
<p style="text-align: center;">Object counting is pivotal for understanding the composition of scenes. Previously, this task was dominated by class-specific methods, which have gradually evolved into more adaptable class-agnostic strategies. However, these strategies come with their own set of limitations, such as the need for manual exemplar input and multiple passes for multiple categories, resulting in significant inefficiencies. This paper introduces a more practical approach enabling simultaneous counting of multiple object categories using an open-vocabulary framework. Our solution, OmniCount, stands out by using semantic and geometric insights (priors) from pre-trained models to count multiple categories of objects as specified by users, all without additional training. OmniCount distinguishes itself by generating precise object masks and leveraging varied interactive prompts via the Segment Anything Model for efficient counting. To evaluate OmniCount, we created the OmniCount-191 benchmark, a first-of-its-kind dataset with multi-label object counts, including points, bounding boxes, and VQA annotations. Our comprehensive evaluation in OmniCount-191, alongside other leading benchmarks, demonstrates OmniCount's exceptional performance, significantly outpacing existing solutions.</p>
<!-- <div style="text-align: center;">
<a href="https://arxiv.org/abs/2403.05435" class="lnk">
<i class="fas fa-file-alt" style="font-size: 24px;"></i>
</a> |
<a href="https://arxiv.org/pdf/2403.05435.pdf" class="lnk">
<i class="fas fa-file-pdf" style="font-size: 24px;"></i>
</a> |
<a href="https://github.com/mondalanindya/OmniCount/tree/main/code" class="lnk">
<i class="fab fa-github" style="font-size: 24px;"></i>
</a>
</div> -->
</div>
<div class="section">
<h2 style="text-align: center;">OmniCount: Model Design</h2>
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/pipeline.png" alt="OmniCount Pipeline" class="full-width">
<p style="text-align: center;">OmniCount Pipeline: Our method starts by processing the input image and their target object classes, using Semantic Estimation and Geometric Estimation modules to generate class-specific masks and depth maps. These initial priors are refined with a Semantic Refinement module for accuracy, creating precise binary masks of target objects. The refined masks help in obtaining RGB patches for each class and also extracting reference points to reduce overcounting. SAM uses these RGB patches and reference points to create instance-level masks, yielding precise object counts. ❄ represents frozen pre-trained models.</p>
</div>
<div class="section">
<h2 style="text-align: center;">Improving Counting using Priors</h2>
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/refinement.png" alt="OmniCount Pipeline" class="full-width">
<p style="text-align: center;">Reference Point Selection: SAM’s segmentation accuracy is enhanced by refining reference point selection. Panel (A) shows how integrating semantic priors, identifying local maxima, and applying Gaussian refinement improve reference point accuracy, focusing them on foreground objects for better segmentation and counting. Panel (B) demonstrates the benefits of incorporating semantic and geometric priors, where depth-based recovery and precise reference points help SAM recover distant or occluded objects, reducing over-segmentation issues found in the default "everything mode"</p>
</div>
<div class="section">
<h2 style="text-align: center;">Results</h2>
<div class="comparison-blocks">
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/agr/agr1.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/agr/agr1_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Potatoes: 4, Apples: 2, Bananas: 3, Onions: 4</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/birds/birds1.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/birds/birds1_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Crow: 9, Pigeons: 10</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/fruits/fruits1.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/fruits/fruits1_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Jackfruit: 1, Lichi: 12, Dragonfruit: 1, Pears: 27, Coconut: 3, Pineapple: 2</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/pets/pets1.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/pets/pets1_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Dog: 1, Cats: 1, Rabbit: 1, Bird: 1, Guineapig: 1, Boar: 1</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/urban/urban1.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/urban/urban1_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Cars: 7</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/house/343.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/house/343_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Strawberries: 16, Kiwis: 14</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/urban/urban3.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/urban/urban3_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Person: 81</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/wild/wild2.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/wild/wild2_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Elephant: 1, Buffaloes: 3</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/urban/london.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/urban/london_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Person: 45</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/agr/pulses.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/agr/pulses_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Beans: 33, Green peas: 169</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/birds/penguin.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/birds/penguin_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Penguins: 197</p>
</div>
<div class="image-block">
<div class="image-compare-container">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/fruits/fruits.jpg" alt="Image 1A">
<div class="overlay">
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/demo/fruits/fruits_mask.png" alt="Image 1B">
</div>
<div class="slider"></div>
</div>
<p class="image-caption">Apples: 79, Peaches: 10</p>
</div>
</div>
</div>
<div class="section">
<h2 style="text-align: center;">Omnicount-191 Benchmark</h2>
<img src="https://raw.githubusercontent.com/mondalanindya/OmniCount/main/assets/figs/omnicount191.png" alt="OmniCount-191 Benchmark" class="full-width">
<p>OmniCount-191: A comprehensive benchmark for multi-label object counting. The dataset consists of 30,230 images with multi-label object counts, including points, bounding boxes, and VQA annotations. For more details, please visit our <a href="https://huggingface.co/datasets/anindyamondal/Omnicount-191" class="lnk">Hugging Face page</a>.</p>
</div>
<div class="section">
<h2 style="text-align: center;">BibTeX</h2>
<center><pre><code>@article{mondal2024omnicount,
title={OmniCount: Multi-label Object Counting with Semantic-Geometric Priors},
author={Mondal, Anindya and Nag, Sauradip and Zhu, Xiatian and Dutta, Anjan},
journal={arXiv preprint arXiv:2403.05435},
year={2024}
}
</code></pre></center>
</div>
<div class="section">
<h2 style="text-align: center;">License</h2>
<p style="text-align: center;">Object counting has legitimate commercial applications in urban planning, event logistics, and consumer behavior analysis. However, said technology concurrently facilitates human surveillance capabilities, which unscrupulous actors may intentionally or unintentionally misappropriate for nefarious purposes. As such, we must exercise reasoned skepticism towards any downstream deployment of our research that enables the monitoring of individuals without proper legal safeguards and ethical constraints. Therefore, in an effort to mitigate foreseeable misuse and uphold principles of privacy and civil liberties, we will hereby release all proprietary source code pursuant to the Open RAIL-S License, which expressly prohibits exploitative applications through robust contractual obligations and liabilities.</p>
</div>
<div class="section">
<!-- <h2 style="text-align: center;">Acknowledgement</h2> -->
<p style="text-align: center; font-size: 10px; ">Inspired by <a href="https://nerfies.github.io/">Nerfies</a>. Thanks to <a href="https://www.linkedin.com/in/manisha-saha-373052341/">Manisha</a> for her exceptional UI/UX design insights.</p>
</div>
</div> <!-- Closing the main container div -->
</div> <!-- Closing the outer container div -->
<script>
document.addEventListener('DOMContentLoaded', () => {
document.querySelectorAll('.image-compare-container').forEach(container => {
const slider = container.querySelector('.slider');
const overlay = container.querySelector('.overlay');
let isDown = false;
slider.addEventListener('mousedown', () => {
isDown = true;
slider.classList.add('active');
});
document.addEventListener('mouseup', () => {
isDown = false;
slider.classList.remove('active');
});
const debounce = (func, wait) => {
let timeout;
return (...args) => {
clearTimeout(timeout);
timeout = setTimeout(() => func.apply(this, args), wait);
};
};
document.addEventListener('mousemove', debounce((e) => {
if (!isDown) return;
const rect = container.getBoundingClientRect();
const x = e.pageX - rect.left;
const walk = Math.max(0, Math.min(x, rect.width));
slider.style.left = `${walk}px`;
overlay.style.clipPath = `inset(0 ${rect.width - walk}px 0 0)`;
}, 10));
});
});
</script>
<center class="noclick">
<a href='https://clustrmaps.com/site/1bt1g' title='Visit tracker'><img src='//clustrmaps.com/map_v2.png?cl=ffffff&w=a&t=n&d=VUWsmjs9vT_QHmhAr6OuY_eMPD1CJyQ5FGORa626Ips&co=37a1ec&ct=ffffff'width="0.003" height="0.002"/></a>
</center>
</div>
</body>
</html>