Closed nies14 closed 3 months ago
Hi @nies14, thanks for your interests. I used the keypoints [25, 24, 23, 18, 19, 20, 4, 3, 5, 7, 16, 15, 14, 9, 10, 11] of the MPI-INF-3DHP dataset. You can directly test using our pretrained model. Btw, this issue might be helpful.
@chaneyddtt Thanks for the reply. I just have 2 queries
Hi @nies14, the MPI-INF-3DHP dataset is pre-processed in the same way as the Human3.6M, namely centralize with the root joint and then normalize. I wrote the pre-precess code following the Human3.6M. After that you can test with the same command as the one we use for Human3.6.
Hi @chaneyddtt Thanks a lot for the reply. Could u please give me your pre-processing script?
If u can't provide that then could you please provide the orginal script link u followed to make the mpi-inf-3dhp pre-processing script?
I did not follow any link since I did not find it back then. I did not clean up the code for this part, so it is a bit messy. Can you give me your email address if you want the script?
@chaneyddtt thanks so much for that
mthnies@gmail.com
I did not follow any link since I did not find it back then. I did not clean up the code for this part, so it is a bit messy. Can you give me your email address if you want the script?
Could u please mail me your preprocess scrips on mpi inf 3dhp? 1499703595@qq.com thanks so muchππ»ππ»ππ»
I did not follow any link since I did not find it back then. I did not clean up the code for this part, so it is a bit messy. Can you give me your email address if you want the script?
Could u please mail me your preprocess scrips on mpi inf 3dhp? 1499703595@qq.com thanks so muchππ»ππ»ππ»
You can follow this repo MLP-Graph
Hi @chaneyddtt Thanks for the great work. I have preprocessed the MPI-INF-3DHP Preprocessed data. I Pre-processed that using following script (https://github.com/nkolot/SPIN/blob/master/datasets/preprocess/mpi_inf_3dhp.py). Now I have 'mpi_inf_3dhp_test.npz'. Now, how can I test on MPI-INF-3DHP using this?
Which keypoints & arguments do I need to pass to test this?