CASIA-IVA-Lab / FastSAM

Fast Segment Anything
GNU Affero General Public License v3.0
7.48k stars 707 forks source link

Doubts about the box detection result, whether it is correct? #86

Open githublqs opened 1 year ago

githublqs commented 1 year ago

image

githublqs commented 1 year ago

This is the final output of the three masks image

zxDeepDiver commented 1 year ago

Hello, is the problem the third mask?

githublqs commented 1 year ago

@zxDeepDiver yes

zxDeepDiver commented 1 year ago

OK, thanks for yout feedback. Currently FastSAM indeed generates these masks for some images, as discussed in our paper. We are working to improve it.

xiangw369 commented 12 months ago

This is using c++?

xiangw369 commented 12 months ago

Can you share your codes?