y0news
← Feed
Back to feed
🧠 AI NeutralImportance 6/10

EgoNight: Towards Egocentric Vision Understanding at Night with a Challenging Benchmark

arXiv – CS AI|Deheng Zhang, Yuqian Fu, Runyi Yang, Yang Miao, Tianwen Qian, Xu Zheng, Guolei Sun, Ajad Chhatkuli, Xuanjing Huang, Yu-Gang Jiang, Luc Van Gool, Danda Pani Paudel||4 views
🤖AI Summary

Researchers introduce EgoNight, the first comprehensive benchmark for nighttime egocentric vision understanding, featuring day-night aligned videos and visual question answering tasks. The benchmark reveals significant performance drops in state-of-the-art multimodal large language models when operating under low-light conditions.

Key Takeaways
  • EgoNight is the first benchmark specifically designed for nighttime egocentric vision understanding with VQA as the core task.
  • The benchmark includes 3658 QA pairs across 90 videos with day-night aligned scenarios to improve annotation quality.
  • State-of-the-art multimodal large language models show substantial performance degradation when transferring from day to night conditions.
  • The dataset combines synthetic Blender-rendered videos and real-world recordings with over 300 hours of human verification work.
  • Two auxiliary tasks include day-night correspondence retrieval and egocentric depth estimation to further test model capabilities.
Read Original →via arXiv – CS AI
Act on this with AI
Stay ahead of the market.
Connect your wallet to an AI agent. It reads balances, proposes swaps and bridges across 15 chains — you keep full control of your keys.
Connect Wallet to AI →How it works
Related Articles