Video-to-Motion Generation

Click the "+" button at the lower left of the chat area to select and upload a local video file. The platform automatically recognizes and extracts the person's motion data from the video and applies it to a 3D character model to reconstruct and generate the motion. Once generation is complete, you can preview the motion in the 3D viewport and use it for subsequent scene arrangement and set previsualization.

Video upload requirements:

• Video Duration: No more than 20 seconds.
• File Size: No more than 50 MB.
• Number of People: Only single-person motion generation is supported.
• Visual Requirements: Keep the person's full body clearly visible and avoid significant occlusion, the person leaving the frame, or frequent shot changes to ensure accurate and continuous motion recognition.

Note: Use video footage with a stable image, complete character movements, and a clearly visible subject for better motion generation results.