MAVE reconstructs editable 3D blockouts from film shots, works inside Blender
bdsqlsz · x · 2026-11-19
bdsqlsz's team released MAVE (Reconstructing Editable 3D Blockouts from Film Shots): a method that recovers camera positions and editable 3D scene blockouts directly from video footage, usable inside Blender via an agent and feeding generation models for novel viewpoints.
The diagrams show the pipeline: metric observations (depth, segmentation, spatial observations, structural measurements) extract geometric evidence; an agentic modeling loop (construct-edit-verify) builds the scene; evidence-based verification checks video consistency, evidence coverage and human-environment consistency.
Related event: MAVE Reconstructs Editable 3D Scenes from Film Shots, Now Open Source(2 posts)→
More from Multimodal
- Scenario open-sources a full AI art team as a Photoshop plugin, MCP-ready — aziz4ai · 2026-11-19
- AI-made Dragon Ball parody music video nails the anime look — Top_Brilliant_9587 · 2026-11-19
- GTA 6-style LoRA generates strikingly authentic rain-soaked neon artwork — freshstart2027 · 2026-11-19
- VAE Rhythm: Detecting AI Videos via Periodic 3D Causal VAE Decoding Traces — showlab · 2026-11-19
- NUS's Jeveration turns a frozen vision-language model into a single-pass image generator — NUS-DSA3101 · 2026-11-19
- KAIST's ReMo: training-free cross-modal token compression trims Omni-LLM inference cost — kaist-ai · 2026-11-19