How to align 2K2K scans with SMPL-X?

[azuxmioy]

Hello,

Thanks for updating the dataset. I've tried the latest SMPL-X parameters in the dataset repo.

I used the following code to create the body mesh

body_model = SMPLX(model_path=os.path.join(MODEL_PATH,'smplx'), use_face_contour=True, use_pca=False, gender='neutral')
smpl_data = json.load(open(os.path.join(smplx_folder, smpl_file)))
output = body_model(  global_orient= torch.tensor(smpl_data['global_orient']).unsqueeze(0).contiguous(),
                          transl = torch.tensor(smpl_data['transl']).unsqueeze(0).contiguous(),
                          body_pose = torch.tensor(smpl_data['body_pose']).unsqueeze(0).contiguous(),
                          betas = torch.tensor(smpl_data['betas']).unsqueeze(0).contiguous(),
                          jaw_pose=torch.tensor(smpl_data['jaw_pose']).unsqueeze(0).contiguous(),
                          left_hand_pose = torch.tensor(smpl_data['left_hand_pose']).unsqueeze(0).contiguous(),
                          right_hand_pose = torch.tensor(smpl_data['right_hand_pose']).unsqueeze(0).contiguous(),
                          expression=torch.tensor(smpl_data['expression']).unsqueeze(0).contiguous(),
                          return_verts=True
                        )

However, there are still translation and rotation offsets between the PLY scans and the SMPLX body mesh.

Also, it seems that the scale (0.986) is not correct. Could you please describe how you registered the body model parameters?

Thank you very much.

[fishfishson]

@azuxmioy Hi, I have met with the same problem. Have you found the solution to this misalignment?

[fengq1a0]

@fishfishson Hi, I also have the same problem. Have you found any solution?
(the given SMPL mesh in the .obj file is 170m tall, however, the mesh in the .ply file is only about 1m)

[fishfishson]

I have no idea. It seems like scans are rescaled for privacy protection.

…

________________________________ 发件人: fengq1a0 ***@***.***> 发送时间: 2024年4月24日 14:15 收件人: SangHunHan92/2K2K ***@***.***> 抄送: YU Zhiyuan ***@***.***>; Mention ***@***.***> 主题: Re: [SangHunHan92/2K2K] How to align 2K2K scans with SMPL-X? (Issue #13) @fishfishson<https://github.com/fishfishson> Hi, I also have the same problem. Have you found any solution? (the given SMPL mesh in the .obj file is 170m tall, however, the mesh in the .ply file is only about 1m) ― Reply to this email directly, view it on GitHub<#13 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AG3KC7WHNANWR3YZJILPYNTY65EXHAVCNFSM6AAAAAA5LTOUJWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANZUGE2DONBWHA>. You are receiving this because you were mentioned.

[yvette256]

Hi, super late to this thread but I'm currently running into a similar issue. I was wondering if anyone managed to figure out the proper alignment?

To the authors, I also wanted to ask if it's possible to clarify whether the rescaling is indeed intentionally for privacy protection, thank you in advance!

[azuxmioy]

Hi all,

This is just an unhelpful update after a year. Since the scale and offset information are missing in the dataset, I ended up registering SMPL-X parameters myself by rendering multi-view images from the scan. But still, the absolute measurement is unknown.

I feel the rescaling is not for privacy protection but a normalization in 3D instead as scans of kids are generally larger. However, the scaling factor and normalization function are not provided, so we don't know the exact measurement of the original human bodies. I'll keep this issue open and see if the authors have any updates.