Chip defect detection based on deep learning method

Abstract

With the rapid development of deep learning theory and computing resources, defect detection based on deep learning has been increasingly used. Compared with traditional machine learning methods, detection methods based on deep learning can achieve end-to-end detection methods, with high flexibility and accuracy, strong network expression capabilities, and no manual design features. This paper focuses on the use of deep learning-based methods to detect chip defects: make data sets according to the types of chip defects, detect chip defect based on the YOLOv3 network and fine-tuning it. The final mAP reached 86.36%.