paTRICK-swk / D3DP

[ICCV2023] The PyTorch implementation for "Diffusion-Based 3D Human Pose Estimation with Multi-Hypothesis Aggregation"
MIT License
158 stars 8 forks source link

Apply D3DP on animal #24

Open Batwho opened 3 months ago

Batwho commented 3 months ago

Hello,

Thanks for sharing this great work. I'm new to the pose lifting area and wonder is it possible to apply D3DP to animal videos. E.g. use a trained D3DP checkpoint and fine-tune on animals' 2d-3d keypoint dataset. I know diffusion-based generally require lots of data to have good performance. Do you think maybe there is still not enough animal data?