How to discard false positive

[chris-aeviator]

When I use P SAM on an image that does not contain my learned item, it will still output a prediction. How can I at least compare a true positive against a false positive?

[vamsidharkamanuru]

You could use thresholding on top confidence keypoint score from similarity mask step. This way when the object you are interested in, is not in the image, the top confidence keypoint may have low score than threshold, resulting in no keypoints selected for prompting decoder.