search

robustsam / RobustSAM

RobustSAM: Segment Anything Robustly on Degraded Images (CVPR 2024 Highlight)
https://robustsam.github.io/
MIT License
152 stars 15 forks source link

I am not getting correct bounding box after inferencing. #3

Closed charchit7 closed 2 months ago

charchit7 commented 2 months ago

I manually checked the bounding box. But while doing the inference with the same bbox. I am getting an differerent bounding box. Could you please tell me why is that the case.

Screenshot 2024-06-26 at 10 56 59 AM results_sofa

robustsam commented 2 months ago

We need to perform transformation on input image in order to match the size requirements of model (please refer to L89 of eval.py), I think that's the cause of your problem.

charchit7 commented 2 months ago

Closing, you are right, I made a mistake. Thank you for your work :)