Blur-to-video: SIGGRAPH Asia 2025 work recovers past, present and future frames from one motion-blurred photo
CSProfKGD · x · 2026-09-12
Presented at the ECCV 2026 doctoral consortium, the generative re-photography project (SIGGRAPH Asia 2025 / ACM TOG, York University, University of Toronto, Vector Institute) reconstructs video from a single motion-blurred image.
- Predicts the present (within exposure time) and extrapolates past and future frames, with frame-level control
- Outperforms MotionETR and Jin et al. on in-the-wild scenes including complex motion and occlusion; predictions feed MegaSAM for dynamic 3D reconstruction
- Applications include bringing historical blurred photos to life and recovering 3D human pose
Paper, code, and demo are available on the project page.
More from Multimodal
- Unlocking any-order generation in pretrained autoregressive image models with future-aware embeddings — Rishav Pramanik · 2026-10-31
- Compositor Open-Sources MatteDesk: Free ComfyUI-Based Shot-Level Alpha Matting Pipeline — fishfinger44 · 2026-10-31
- Gemini 4 Carbon vs Opus 5.5 hands-on: 3x cheaper at $4.10 and twice as fast — vitaliychiley · 2026-10-31
- Surreal AI fairytale: a town obsessed with turning tears into yarn — gizakdag · 2026-10-31
- ComfyUI extension adds one-click saving of preview images with metadata — sitefall · 2026-10-31
- DeepMind scientist directs MKUltra microdrama with Gemini agents for ~$35 total — korymath · 2026-10-31