Pytorch Implement of the paper "Segmentation Based Deep-Learning Approach for Surface DefectDetection" https://arxiv.org/abs/1903.08536
Modified little things without changing the accuracy
BUT, Much Faster than the paper!!!
Achieve less 3ms per image of the size 704x256, by a single Nvidia 1080Ti. This speed can rival Cognex's VIDI ~
step1: train segment net
python train_segment.py
step2: train decision net
python train_decision.py
step3: test
python test.py