Search code examples
deep-learningconv-neural-networkyolonon-maximum-suppression

YOLO - does Intersection over Union is actually a part of Non Maximum Suppresion

Here:[related topic] 1 they say that in YOLO actually the IoU (intersection over union) used twice:

1- during training to compare ground truth box to predicted box

2- during the usage of already trained YOLO network this technique is being used to eliminate overlapping boxes which include same object many times.

As far as I know eliminating overlapping boxes is being done by process called Non Maximum Suppression (NMS). Thus I wonder maybe the IOU is a part of a NMS process?

Solution

  • Non Maximum Suppression eliminates overlapping boxes by dropping predicted bounding boxes that have high IOU with eachother.

    Lets say two predicted bounding boxes box1 and box2 have an IOU of 99%. This means that both boxes are predicting the same object, thus one of them should be dropped.