Six ECCV papers from one team: multimodal, agents, 3D and embodied AI
liuziwei7 · x · 2026-09-08
liuziwei7 shares his team's ECCV 2025 accepted papers:
- Multimodal & agents: Prism Hypothesis and HippoCamp (hippocampus-inspired agent memory)
- 3D & embodied AI: HSImul3R, MonoArt (monocular 3D generation), and InfiniteDance (dance motion generation)
All papers include project/GitHub links, spanning multimodal understanding to embodied motion generation.
More from Multimodal
- Encoding entire images into audio spectrograms with Gemini's Astra — seanmcdonaldxyz · 2026-09-08
- Hailuo H3 recreates World Labs Atlas-style spatial video in under 7 seconds — Hailuo_AI · 2026-09-08
- Seedance 2.5 Prompt Recreates Uncannily Real Early-2000s DV Home Video — SimplyAnnisa · 2026-09-08
- JSON Prompt Restores Old Photos With Damage Repair and 4x ESRGAN Upscaling — HeyZoyaKhan · 2026-09-08
- Creator animates childhood stamp collection with Hailuo AI and Codex — Hailuo_AI · 2026-09-08
- Hajimete3D ships image-to-3D MCP workflow where agents approve a quote before paying for generation — Shoddy-Technology950 · 2026-09-08