Following repository consists code to convert the ground truth SMPL-X files (neutral) provided on AGORA website in a format use for training Human pose and shape estimiation methods. One could modify the code to generate output in desire format for training.
Run following script to download data. Note that you need to register on SMPL-X and AGORA website to get access to the data
bash fetch_data.sh
Alternatively, you could manually dowload the files from the website. Please note that all the downloaded images should be combined in single folder data/images
python dataframe_agora_smplx_release.py
imgname = Name of the images
center = center of the subject
scale = scale of the subject
pose_cam = SMPL-X pose parameter of the subject in camera coordinates
pose_world = SMPL-X pose parameter of the subject in world coordinates
shape = 11 SMPL-X shape componenets
gtkps = 2d joints
cam_int = camera intrinsic matrix
cam_ext = camera extrinsic matrix
proj_verts = 2d vertices subsampled
bfh = body or bfh