Research map / Video / Video generation
Sketch2Colab: Sketch-Conditioned Multi-Human Animation via Controllable Flow Distillation
Divyanshu Daiya, Aniket Bera
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add energy guidance (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | Trajectory-3D error: .149 -> .116 | Improved |
| add energy guidance (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | Anchor error: .075 -> .038 | Improved |
| add energy guidance (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | FID: .471 -> .399 | Improved |
| add CTMC phase scheduling (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | Anchor error: .050 -> .038 | Improved |
| add CTMC phase scheduling (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | Penetration: .027 -> .022 | Improved |
| add COLLAGE latent grounding (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | Anchor error: .047 -> .038 | Improved |
| add temporal attention bias (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | Trajectory-3D error: .122 -> .116 | Improved |
| use unified conditioning path instead of parallel ControlNets (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | FID: .440 -> .399 | Improved |
| use unified conditioning instead of trajectory-only route (ablation) | Sketch2Colab · CORE4D sketch+text (HOH storyboard) | Anchor error: .061 -> .038 | Improved |
| distill rectified-flow student with guidance instead of COLLAGE diffusion teacher | COLLAGE Teacher · CORE4D sketch+text (HOH storyboard) | FID: .511 -> .399 | Improved |
| distill rectified-flow student with guidance instead of COLLAGE diffusion teacher | COLLAGE Teacher · CORE4D sketch-only (HOH storyboard) | FID: .703 -> .487 | Improved |
| use Sketch2Colab instead of retrieval-based RetrievAdapt | RetrievAdapt · CORE4D sketch+text (HOH storyboard) | Penetration: .019 -> .022 | Hurt |
Closest papers by meaning
- Human Motion Diffusion as a Generative Prior (ICLR 2024 · similarity 0.87)
- Enhanced Fine-Grained Motion Diffusion for Text-Driven Human Motion Synthesis (AAAI 2024 · similarity 0.87)
- StickMotion: Generating 3D Human Motions by Drawing a Stickman (CVPR 2025 · similarity 0.87)
- FlipSketch: Flipping Static Drawings to Text-Guided Sketch Animations (CVPR 2025 · similarity 0.86)
- It's All About Your Sketch: Democratising Sketch Control in Diffusion Models (CVPR 2024 · similarity 0.86)
- AirSketch: Generative Motion to Sketch (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
