Segnet outputs mask even though there is no object in frames
Closed this issue · 1 comments
TrinhNC commented
Hi,
I trained SegNet on my own dataset, and it worked well if the input frames including target object. But when there are no objects in frames, SegNet still outputs masked pixels. Does anyone know how to avoid this? I also tried Unet but same things happened.
j96w commented
Hi, could you double check whether your segnet has a 0-channel for the background. If so, you should not have this issue.