Negative/no object data for YoloV4


I want to train the YOLOV4 detctor with images of set background, so when there is no object, it doesn’t create BBox on the background. What is the best way to do this?

Please provide the following information when requesting support.

• Hardware (3090)
• Network Type (Yolo_v4)
• TLT Version (TAO 21.08)

When prepare the training dataset, it is needed to use images which have labels.
And after training, then do inference, actually when there is no object, it doesn’t create any BBox during inference.

The problem is that it creats BBox during inference, when there is no object. Trying to find a solution for it?!

How about the mAP?

There is no update from you for a period, assuming this is not an issue anymore.
Hence we are closing this topic. If need further support, please open a new one.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.