una-dinosauria/3d-pose-baseline

Question about output p3d and allposes3d

Closed this issue · 2 comments

Hi , I have two questions ,

  1. I have had 2D detections which i feed it into the 3d predictor. I tried to trace the output of the following two lists poses3d and all_poses_3d . Regarding the output of all_poses_3d which is as follows. Is it the 3D predictions (X,Y,Z) from the entire video sequence ??? Does each row correspond to frame or is it time ?

[array([[ 0.00000000e+00,  0.00000000e+00,  0.00000000e+00,
        -1.18590513e+02, -7.54710593e+01, -4.81356906e+01,
        -5.67777077e+01,  9.05888615e+01, -1.30586482e+02,
        -2.60462532e+02,  3.60059093e+02,  1.70836726e+01,
        -1.15560633e+01,  7.42149725e+02,  1.66477287e+02,
        -1.18447102e+01,  7.36763064e+02,  1.65182437e+02,
         1.18589863e+02,  7.54705369e+01,  4.81355084e+01,
         9.51023920e+01,  4.36000692e+02,  2.89570966e+01,
        -4.54617394e+01,  8.14331952e+02,  1.65912608e+02,
        -1.12642400e+01,  7.48636864e+02,  1.66665977e+02,
        -1.14090840e+01,  7.36435064e+02,  1.63713810e+02,
         1.21660845e-03, -8.60110629e-02, -1.93000576e-02,
         4.83018609e+01, -2.11079417e+02, -4.82572435e+00,
         1.39187234e+02, -4.19865990e+02, -7.02882449e+01,
         1.69620093e+02, -4.50983988e+02, -1.74176445e+02,
         1.34253334e+02, -5.62355770e+02, -1.57637814e+02,
         5.67537804e+00, -4.35088906e+02, -9.76974016e+01,
         2.14618565e+02, -3.30512564e+02, -1.09704604e+01,
         3.57793036e+01, -1.69992728e+02,  5.44005206e+01,
         1.92711691e+02, -8.23625625e+01,  1.02046721e+02,
         1.26585822e+00, -1.20170579e+02, -2.82526049e+01,
         1.57900698e+00, -1.51780249e+02, -3.52080548e+01,
         8.84543993e-01, -1.07795356e+02, -2.56307189e+01,
         8.84543993e-01, -1.07795356e+02, -2.56307189e+01,
         5.67537804e+00, -4.35088906e+02, -9.76974016e+01,
         2.98958555e+01, -4.21684909e+02, -1.03864265e+02,
         1.93760533e+01, -2.89760642e+02, -5.48227057e+01,
        -4.60343417e+01, -2.60396688e+02, -7.98165678e+01,
         8.70569014e-01, -1.68664569e+02, -3.73902498e+01,
         1.39982512e+00, -2.00884252e+02, -4.47207875e+01,
         5.24591115e-01, -1.65867774e+02, -3.68342864e+01,
         5.24591115e-01, -1.65867774e+02, -3.68342864e+01]])]


And poses3d what does it output to ? It seems to have 4 coloumns . The array list is below ( I guess this is also for the entire video sequence ?? ) Sequence output below

[[-1.88248140e+02  0.00000000e+00  6.20255861e+02 -3.06838653e+02
  -4.81356906e+01  6.95726921e+02 -2.45025848e+02 -1.30586482e+02
   5.29667000e+02 -4.48710673e+02  1.70836726e+01  2.60196769e+02
  -1.99804203e+02  1.66477287e+02 -1.21893864e+02 -2.00092850e+02
   1.65182437e+02 -1.16507202e+02 -6.96582775e+01  4.81355084e+01
   5.44785324e+02 -9.31457481e+01  2.89570966e+01  1.84255170e+02
  -2.33709879e+02  1.65912608e+02 -1.94076091e+02 -1.99512380e+02
   1.66665977e+02 -1.28381002e+02 -1.99657224e+02  1.63713810e+02
  -1.16179203e+02 -1.88246923e+02 -1.93000576e-02  6.20341872e+02
  -1.39946279e+02 -4.82572435e+00  8.31335278e+02 -4.90609059e+01
  -7.02882449e+01  1.04012185e+03 -1.86280467e+01 -1.74176445e+02
   1.07123985e+03 -5.39948062e+01 -1.57637814e+02  1.18261163e+03
  -1.82572762e+02 -9.76974016e+01  1.05534477e+03  2.63704247e+01
  -1.09704604e+01  9.50768425e+02 -1.52468836e+02  5.44005206e+01
   7.90248589e+02  4.46355101e+00  1.02046721e+02  7.02618424e+02
  -1.86982282e+02 -2.82526049e+01  7.40426440e+02 -1.86669133e+02
  -3.52080548e+01  7.72036110e+02 -1.87363596e+02 -2.56307189e+01
   7.28051217e+02 -1.87363596e+02 -2.56307189e+01  7.28051217e+02
  -1.82572762e+02 -9.76974016e+01  1.05534477e+03 -1.58352285e+02
  -1.03864265e+02  1.04194077e+03 -1.68872087e+02 -5.48227057e+01
   9.10016503e+02 -2.34282482e+02 -7.98165678e+01  8.80652549e+02
  -1.87377571e+02 -3.73902498e+01  7.88920430e+02 -1.86848315e+02
  -4.47207875e+01  8.21140114e+02 -1.87723549e+02 -3.68342864e+01
   7.86123635e+02 -1.87723549e+02 -3.68342864e+01  7.86123635e+02]]
  1. Has anyone managed to export the 3d skeletal to unity aavtar ? I was trying to follow one of the links i saw in the discussions before but could really get any sense of it
    [https://qiita.com/romaroma/items/ffbdae4ecfc4c8ff31cd]

I suggest you try out our visualization function:

def show3Dpose(channels, ax, lcolor="#3498db", rcolor="#e74c3c", add_labels=False): # blue, orange
"""
Visualize a 3d skeleton
Args
channels: 96x1 vector. The pose to plot.
ax: matplotlib 3d axis to draw on
lcolor: color for left part of the body
rcolor: color for right part of the body
add_labels: whether to add coordinate labels
Returns
Nothing. Draws on ax.
"""

to understand the output format.

Closing for lack of activity. Please reopen if the issue is still ongoing.