Explore the Potential of CLIP for Training-Free Open Vocabulary Semantic Segmentation.

Tong Shao, Zhuotao Tian, Hang Zhao, Jingyong Su

Browse the full ECCV paper archive.