nodefluxio/face-detector-benchmark

precision calculation error

Opened this issue · 0 comments

precision = float(tp) / float(total_gt_face)

In that line true positives are divided by the number of ground truths. Shouldn't the precision be equal to true positives divided by the number of predictions inside an image, hence;precision = true positives / (true positives + false positives)?

precision = float(tp) / float(len(pred_dict.keys()))