Open Samyk01 opened 1 year ago
Hi, could you please describe the use case? I'm not sure segmentation models can work with pixel precision.
Hi, could you please describe the use case? I'm not sure segmentation models can work with pixel precision.
It's for lesion segmentation on medical images. Some lesions are quite small, when i draw a polygon around them and press N the polygon just disappears, i need to make the shape bigger for it to stay.
As a workaround you can try using points. Probably this restriction can be redesigned.
When drawing polygons the lowest size it can go is 3x3 pixels, i need to segment details that are sometimes 2x1 pixels, is it not possible on cvat to do so ?