From Filters to VLMs: Benchmarking Defogging Methods through Object Detection and Segmentation Performance
从滤波器到视觉语言模型:通过目标检测和分割性能评估去雾方法
机构 * University of Southern California(南加州大学)
专题命中 视觉定位与Grounding :visual language model(abstract);分类 cs.CV;VLM(comments)
AI总结 本文通过目标检测和分割性能评估,探讨了去雾方法在真实与合成环境中的有效性,揭示了视觉语言模型在恶劣天气下的应用潜力。
Comments Accepted at WACV 2026 Proceedings (Oral), 5th Workshop on Image, Video, and Audio Quality Assessment in Computer Vision, with a focus on VLM and Diffusion Models