Learning to See through Illumination Extremes with Event Streaming in Multimodal Large Language Models
通过事件流学习在极端光照下视觉感知
机构 * The University of Hong Kong(香港大学)
专题命中 视觉推理 :multimodal large language model(title,abstract);visual reasoning(abstract);MLLM(abstract);分类 cs.CV
AI总结 本文提出Event-MLLM,通过动态融合事件流与RGB帧进行全光视觉推理,引入光照指示器和光照校正损失,解决极端光照下多模态大语言模型的感知与推理问题。
Comments IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) 2026