search

vicoslab / mixed-segdec-net-comind2021

Official PyTorch implementation for "Mixed supervision for surface-defect detection: from weakly to fully supervised learning"
Other
291 stars 89 forks source link

It seems no evalutation about segmentation #5

Closed HustHB closed 3 years ago

HustHB commented 3 years ago

First, thanks for your public nice code!

I just wonder whether the AP metric in paper is about classification rather than semantic segmentation. And in your "end2end.py " L246, it seems only evaluate for "predictions" which is result of classification.

By the way, I think "weakly supervision" is not precise in your mix supervision setting, because if your final evaluation is only about classification, class-level, pixel-level labels will all be full-supervised label for classification.

JakobBozic commented 3 years ago

Hi, all of the metrics that we report in the paper are about classification, as it is the most relevant metric in industrial defect detection.

smiler96 commented 3 years ago

Hi, all of the metrics that we report in the paper are about classification, as it is the most relevant metric in industrial defect detection.

The AP is adopted the metric for evaluation, would you provide the AUC values both in this repo?

HustHB commented 3 years ago

Hi, all of the metrics that we report in the paper are about classification, as it is the most relevant metric in industrial defect detection.

Thanks for your rapid reply! ^_^