Gen2Balance at ECCV 2026: generative videos fill long-tailed action datasets to reach SOTA
dimadamen · x · 2026-09-10
Presented at ECCV 2026 (poster #19, session S2, 4:30PM), Gen2Balance: Generative Balancing for Long-Tailed Video Action Recognition asks whether imbalanced action recognition datasets can be filled with generative videos to achieve SOTA — and shows that they can.
More from Multimodal
- An interactive journey through the 3D ecosystem: why now is the best time to start — SergioPaniego · 2026-09-10
- MiniMax Design launches: agent-driven studio pairs GPT Image 2.5 with H3 for end-to-end ads — HeyNayeem · 2026-09-10
- Recreating the iPhone Duo ad with GPT Image 2.5 and MiniMax H3 — HeyNayeem · 2026-09-10
- Batman TAS fan scene remade locally with MiniMax H3 on a 16GB 4070 Ti Super — darthfurbyyoutube · 2026-09-10
- KIE Nodes Next turns KIE.ai's API catalog into a native ComfyUI model library — felipederosilva · 2026-09-10
- User reverse-engineers a movie look book by feeding Tom Cruise's 'Digger' to ChatGPT — justin_hart · 2026-09-10