Keypoint names: according to what output format ?

[HabibaAbderrahim]

Hello , after the inference using this command: python demo/inferencer_demo.py tests/data/segbig/img.jpg --pose2d yoloxpose_s_8xb32-300e_coco-640.py --pose2d-weights yoloxpose_s_8xb32-300e_coco-640-56c79c1f_20230829.pth --pose3d human3d --vis-out-dir vis_results/human3d/3_person_4_segmentationt.jpg --device cpu
and printing the 17 keypoints as follow :

I wanted to make sure what each point represent in order to properly map it ( i mean for example :
nose: [-0.0, 0.0, 0.6314282417297363]
left_eye: [0.10871679335832596, 0.07088106125593185, 0.6354875564575195]
right_eye: [0.1409938931465149, -0.30314338207244873, 0.3504978120326996] ?

[HabibaAbderrahim]

i just want to know the printed keypoints representing the Human3.6 keypoints in order like
| 1 | Nose |
| 2 | Neck |
| 3 | Right Shoulder |
| 4 | Right Elbow |
| 5 | Right Wrist |
or representing the coco keypoints in order like
1 & Nose
2 & Left Eye
3 & Right Eye
4 & Left Ear
5 & Right Ear

thank you ,

[Ben-Louis]

Hi @HabibaAbderrahim, the outputs of 3D models follow the keypoints definition in Human3.6M

[HabibaAbderrahim]

Thank you for your reply ,