Data mismatch for MPI_INF_3DHP dataset

[penincillin]

Hi, I am process the MPI_INF_3DHP dataset for training. I found that the number of data samples I obtained mismatches with the static_fits you provided.

After preprocessing, I got 96719 samples for MPI_INF_3DHP dataset while the number of sample included in "mpi-inf-3dhp_mview_fits.npz" is 96507.

I am wondering would you mind provide a list of training images of MPI_INF_3DHP corresponding to the static fits file ?

Thanks !

[penincillin]

@geopavlakos @nkolot @ikvision Would you mind provide me the list of images of MPI_INF_3DHP you used ?

Thanks !

[HongwenZhang]

Hi, I think the image name list of MPI_INF_3DHP can be found in data/dataset_extras/mpi_inf_3dhp_train.npz using the key imgname. The length of mpi_inf_3dhp_train.npz is matched with mpi-inf-3dhp_mview_fits.npz.

[bragilee]

Got the same problem when I tried to process this dataset. The number did not match.

[penincillin]

Hi, I think the image name list of MPI_INF_3DHP can be found in data/dataset_extras/mpi_inf_3dhp_train.npz using the key imgname. The length of mpi_inf_3dhp_train.npz is matched with mpi-inf-3dhp_mview_fits.npz.

Aha, that's it. Thanks for your reminder !

[GloryyrolG]

I guess some selection may be performed, since I found S1/Seq1/imageFrames/video_0/frame_000619.jpg follows S1/Seq1/imageFrames/video_0/frame_000481.jpg, and S1/Seq1/imageFrames/video_0/frame_000859.jpg is followed by S1/Seq1/imageFrames/video_0/frame_001377.jpg, etc. Hi @nkolot , could you help check this? Thanks & regards,