I know the bbox information of BIWI are generated by a MTCNN detector. but it can not detect all face in BIWI datasets.
In my project, MTCNN can only detect 15109/15678. So I want to ask you about how to process the images which doesn't have a face bbox information? Discard them or manual labeling them?
thanks for your useful code!
I know the bbox information of BIWI are generated by a MTCNN detector. but it can not detect all face in BIWI datasets. In my project, MTCNN can only detect 15109/15678. So I want to ask you about how to process the images which doesn't have a face bbox information? Discard them or manual labeling them?