search

rongakowang / DenseMutualAttention

[WACV 2023] Interacting Hand-Object Pose Estimation via Dense Mutual Attention
MIT License
35 stars 3 forks source link

how to render_mesh #12

Closed zy123456123 closed 5 months ago

zy123456123 commented 5 months ago

Hi @rongakowang,

Thanks for sharing the code. How to visualize the results？

rongakowang commented 5 months ago

Hi,

You can extract the predicted vertices and corresponding MANO and object faces to save it as a mesh. All predicted are in the camera frame hence you can directly render the result using external libraries, e.g. pytorch3d or directly view the mesh object.

zy123456123 commented 5 months ago

Thank you very much for your reply. When I use the render_mesh method in your code, the rendering result is not very good.Why are they all rendered white? ![Uploading 2024-04-11 20-01-04屏幕截图.png…]() ![Uploading 图片.png…]()

rongakowang commented 5 months ago

Hi,

I believe the render_mesh you refer to are left by the author in the handAR repo (not implemented by us) and has been deprecated. I would suggest saving the hand and object meshes first to visualize. If the image is full white, it might be that the camera parameters are not set correctly.

zy123456123 commented 5 months ago

ok! Thank you very much for your reply!