Open ImSuMyatNoe opened 1 year ago
Haha, I have the same problem as you. We need to segment a specific category without conducting special training for this specific category (in this case, just use an advanced segmentation model). Before seeing SAM, I thought this requirement was baffling.
Is there any solution for this?
In SAM, when utilizing YOLOv8, it is possible to segment a specific desired class from the COCO class dataset, rather than segmenting everything. However, when working with FASTSAM, I encountered a challenge in filtering out only one class. My goal is to avoid segmenting the background and instead, I want to segment all instances of a particular class, such as dogs from the COCO dataset. I'm looking for guidance on how to achieve this, as I haven't come across similar issues in my research. Can someone provide an explanation or solution?