The pose format in mv3dpose.py

[ligaoqi2]

Amazing work! I found [[pose1, ...]] * frames in tracking.py, but when I run before the 86 lines in baseline.py, I don't know the pose format in "for pose in loc_pred:". If choose current frame in line 49 "for t in range(n_frames):" in tracking.py, why use for pose in pic_pred? Could you provide the format and meaning of the loc_pred and the pose format of the frame00000000m.json? Thank you so much~

[ligaoqi2]

Another question is the "J = len(points2d)" in line 284 hypothesis.py, what's the meaning of the J? Is it the joints number in frame t?

[ligaoqi2]

I also want to ask how many persons is needed in this repo, I found hyp.get_3d_person() needs two person at least.

[ligaoqi2]

I want to do single person's 3D pose reconstruction but I can't matching two pictures 2D pose, Could you give me some suggestions?