Search Results for author:
Found 2 papers, 0 papers with code
Segment Anything Meets Semantic Communication
In light of the diminishing returns of traditional methods for enhancing transmission rates, the domain of semantic communication presents promising new frontiers.
Understanding Segment Anything Model: SAM is Biased Towards Texture Rather than Shape
Different from label-oriented recognition tasks, the SAM is trained to predict a mask for covering the object shape based on a promt.
Cannot find the paper you are looking for? You can
Submit a new open access paper.
