How to apply SMPL model (24 3Dkeypoints) not AMASS (22 3Dkeypoints) ?

[dongran]

Thank you for providing a great tool for calculating IK!

Can I apply the SMPL model?

Because the demo was only for AMASS model 3Dkeypoints.

I have tried the following mapping, but it does not work.

Did you already provide a method to apply SMPL model in the code?

`SMPL_JOINT_NAMES = [
"pelvis",
"left_hip",
"right_hip",
"spine1",
"left_knee",
"right_knee",
"spine2",
"left_ankle",
"right_ankle",
"spine3",
"left_foot",
"right_foot",
"neck",
"left_collar",
"right_collar",
"head",
"left_shoulder",
"right_shoulder",
"left_elbow",
"right_elbow",
"left_wrist",
"right_wrist",
"left_hand",
"right_hand",
]

#According to SMPL keypoints3d method
#################################################
smpl = None
smpl = SMPL(model_path=args.smpl_dir, gender='MALE', batch_size=1)
keypoints3d = smpl.forward(
global_orient=torch.from_numpy(smpl_poses[:, 0:1]).float(),
body_pose=torch.from_numpy(smpl_poses[:, 1:]).float(),
transl=torch.from_numpy(smpl_trans / smpl_scaling).float(),
).joints.detach().numpy()[:, 0:24, :]
nframes = keypoints3d.shape[0]
dances.append(keypoints3d.reshape(nframes, -1).tolist())
print(np.shape(dances[-1])) # (nframes, 72)
################################################

AMASS_JOINT_NAMES = [
"pelvis",
"left_hip",
"left_knee",
"left_ankle",
"left_foot",
"right_hip",
"right_knee",
"right_ankle",
"right_foot",
"left_shoulder",
"left_elbow",
"left_wrist",
"right_shoulder",
"right_elbow",
"right_wrist",
"spine1",
"spine2",
"spine3",
"neck",
"head",
"left_collar",
"right_collar",
]

#According to Amass keypoints3d provided in the config.py

`