Dataset for surface defect detection and prediction in laser processed cutting tools using machine learning
This datasets relates to research for surface defect detection from laser processing in cutting tool materials. This data contains an example image data set of 78 SEM microstructure images processed at various laser parameters. The MATLAB binary data compressed file (.mat) is the corresponding ground truth information, which categorizes the main types of defects. Detection network (.xlsx) - This data file details the AlexNet R-CNN network, training progress via transfer learning, and the detection result. Predictive network (.xlsx) - This data file details the BPNN network, training data from R-CNN output, and the validation performance.