Avant-Gardiste
commented
4 years ago Don’t be fooled about the ‘confidence’ name, as lower confidences are better because it means the distance between the two histograms is closer. We use a threshold and the ‘confidence’ to automatically estimate if the algorithm has correctly recognized the image. We can assume that the algorithm has successfully recognized if the confidence is lower than the threshold defined.
Your work is simply amazing. During my implementation, to leverage the model confidence value I'm using 100 images and still, the algorithm predicts the confidence value of less than 50. In addition, I observe many false positives during recognition. I tried tweaking the confidence values but still, the same problem persists. Am I missing something here?