Paper

Fuse to Recognize the Unseen: Class-Relative Infrared-Visible Fusion for Open-Vocabulary Semantic Segmentation · ByteTrail