.png){kind=icon}
.png){kind=icon}
Return
Long-range diffusion for weakly camouflaged object segmentation
.png){kind=icon}
PRE
AI
DOI：10.1016/j.neunet.2025.107915.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We design a novel long-range diffusion network for WSCOS. • The GLSC loss is designed to take full advantage of sparse annotations. • The Trans-decorator and RUp modules are designed to capture the long-range dependency. • LRDNet has good performance and generalization with points and scribbles.
Keywords:
long-range diffusion network
GLSC loss
Trans-decorator
RUp module
weakly supervised semantic segmentation
Journal
IF:
6.3
Papers:
7.8K
Citations:
3.0W