Tencent / MimicMotion

High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance
https://tencent.github.io/MimicMotion/
Other
1.93k stars 165 forks source link

Question about the inference #29

Open jby1993 opened 4 months ago

jby1993 commented 4 months ago

I noticed that the pose estimation code requires the first frame's pose image to be the reference pose image in each segment. Why is this setting in place? Typically, it should be sufficient to infer from a reference image and a sequence of consecutive pose images, without requiring the first frame's pose image to serve as the reference image for each segment.