UNC postdoc lands 5 papers at ECCV 2026 spanning world generation and video reasoning
mohitban47 · x · 2026-09-08
A postdoc at UNC announced 5 papers accepted to ECCV 2026, covering consistent world generation, error-driven 3D grounding, grounded video reasoning, physical plausibility evaluation, and visual representation learning. Collaborators will be at the conference to discuss the work.
More from Multimodal
- Seedance 2.5 turns a single 30s prompt into a cinematic solar system documentary trailer — LudovicCreator · 2026-09-08
- Text-to-Video Prompting Is a Transition: The Coming Text-Model-to-Video-Model Pipeline — karminski3 · 2026-09-08
- Fable 5 model crafts kinetic typography poem about its own transformation — repligate · 2026-09-08
- gazeCOM: Open-Source Gaze- and VLM-Driven Outpainting on Top of ComfyUI — grebenshyo · 2026-09-08
- mmh3_media: ComfyUI Toolkit Stitches MiniMax H3 Videos via Latent Splicing — EinhornArt · 2026-09-08
- Krea Launches Creative-Work Agents, Calling Them the Future of Media Generation — TitusTeatus · 2026-09-08