yolo2-pytorch ground truth value for iou

ground truth value for iou

Open JesseYang opened this issue 7 years ago • 5 comments

In the code, the ground truth value for iou prediction is the iou between the anchor and the ground truth box (https://github.com/cory8249/yolo2-pytorch/blob/master/darknet.py#L111). But the paper says that " the objectness prediction still predicts the IOU of the ground truth and the proposed box ...". In my understanding the "proposed box" here is the predicted box, not the anchor box.

May 05 '17 07:05 JesseYang

Thank you for pointing out this. I read the darknet code again and think that you are right. The darknet uses anchors to match gt_boxes and pred_boxes (https://github.com/pjreddie/darknet/blob/master/src/region_layer.c#L264) but uses ious between gt_boxes and pred_boxes as targets (https://github.com/pjreddie/darknet/blob/master/src/region_layer.c#L280).

I will fix this and try to train the network later.