Open
Description
How does uio distinguish between instances of different categories during segmentaion training?My understanding is: when uio is training for segmentation, each sample will only require one category to be segmented, and only need to use randomly generated colors to represent different instances. However, it cannot handle the situation where a picture has multiple category requests
Metadata
Metadata
Assignees
Labels
No labels