Automatic Microstructure Image Analysis for Predicting Metals Using Deep Learning

This paper presents a deep learning-based approach for automated microstructure classification in metallic materials. The model leverages convolutional neural networks (CNNs) with transfer learning to enhance accuracy and scalability, achieving a classification accuracy of 95%. Data augmentation was applied to address dataset constraints. This research provides a significant advancement in industrial applications such as aerospace and manufacturing, streamlining metallurgical analysis while improving reliability and decision-making.