Question about Zero-shot image tasks

[1170300714]

Thanks for your great work!

I have a question about the zero-shot image tasks.

Since the SAM-2 will predict three mask for each-time forward, which mask was used to calculate the mIoU in the table results (\eg, Tab 6 )? By default, is the mask with the highest score selected?

Looking forward to your reply~ Thanks a lot

[ronghanghu]

Since the SAM-2 will predict three mask for each-time forward, which mask was used to calculate the mIoU in the table results (\eg, Tab 6 )? By default, is the mask with the highest score selected?

Yes, the predicted mask with the highest predicted IoU score is selected.