Abstract: Referring Image Segmentation (RIS) aims to accurately match specific instance objects in an input image with natural language expressions and generate corresponding pixel-level segmentation ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback