3 accepted papers
We introduce Daytime-Memory Guided Nighttime Image Enhancement (DMGNIE) framework, the first framework that turns long-running daytime surveillance videos of a single intersection into persistent “daytime memory” to guide nighttime image enhancement in traffic scenes. Our key insight is simple yet p
Large Language Models (LLMs) demonstrate strong capabilities in broad knowledge representation, yet they are inherently deficient in pixel-level perceptual understanding. Although the Segment Anything Model (SAM) represents a significant advancement in visual-prompt-driven image segmentation, it exh
Open-domain dialogue systems aim to generate natural and engaging conversations, providing significant practical value in real applications such as social robotics and personal assistants. The advent of large language models (LLMs) has greatly advanced this field by improving context understanding a