Machine Learning Based SEM Image Analysis for Automatic Detection and Classification of Wafer Defects

A machine learning (ML) based approach is proposed for analyzing Scanning Electron Microscope (SEM) images and classifying review defects. Accurate and timely SEM image analysis is crucial and impacts manufacturing yield. The state-of-the-art object detection model YOLOv8 (You Only Look Once version 8) is used as it offers a good balance between accuracy and inference speed. This work demonstrates the utility of YOLO for SEM ADC and extends capability using ensemble voting to achieve higher quality results.