DefectSegNet is a steel defect detection model introducted in the paper Deep Learning for Semantic Segmentation of Defects in Advanced STEM Images of Steels. The implementation by the paper's authors is written in tensorflow and keras. This repository is aimed at assisting those who wish to implement DefectSegNet in pytorch. The model in defectsegnet.py is built to closely resemble DefectSegNet. There may be some features used in the original DefectSegNet (such as Dropout) that are not implemented in this repository. But, they can be added simply by making changes to the DenseConvBlock. Further details of this implementation are provided in the python notebook. This repository does not contain the pre-trained weights of the model.