🎬 Video to Point Cloud Renderer + TTM Video Generation

Upload a video to generate a 3D point cloud, render it from a new camera perspective, and optionally run Time-to-Move (TTM) for motion-controlled video generation.

Workflow:

  1. Step 1: Upload a video and select camera movement → Generate motion signal & mask
  2. Step 2: (Optional) Run TTM to generate a high-quality video with the motion signal

TTM (Time-to-Move) uses dual-clock denoising to guide video generation using:

  • first_frame.png: Starting frame
  • motion_signal.mp4: Warped video showing desired motion
  • mask.mp4: Binary mask for motion regions

📥 Input
🎥 Camera Movement

Select how the camera should move in the rendered video

Generate motion_signal.mp4 and mask.mp4 for Time-to-Move

📤 Rendered Output

🎯 TTM: Motion Signal

🎭 TTM: Mask

Ready to process...

📁 Examples
Examples
Upload Video 🎥 Camera Movement 🎯 Generate TTM Inputs