How do the parameters "ignore_thresh" and "iou_thresh" contribute to the training of a yolo detector model ?

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )

http://pjreddie.com/darknet/

Other

21.56k stars 7.94k forks source link

How do the parameters "ignore_thresh" and "iou_thresh" contribute to the training of a yolo detector model ? #8899

Open riririririi opened 2 months ago

riririririi commented 2 months ago

Hi, I am a little confused about the parameters "iou_thresh" and "ignore_thresh". As per my understanding, "iou_thresh" decides if the predicted bbox is to be considered during training to contribute to the loss if the iou of predicted bbox > ground truth bbox. I feel like "ignore_thresh does" the same thing but i am unable to understand it clearly.