Abstract
Manual visual inspection process is the most used method for performing visual inspection in the manufacturing of semiconductors electronics. It is a critical quality management procedure that all leading automotive manufacturers continue to perform for defect detection. However, manual inspections can be a slow process and increases the chances for error. Striving for lower error rates requires investment in skilled, experienced, and well-trained inspectors. Adding automation and AI (Artificial Intelligence) capabilities to the manual visual inspection process can help reduce human visual perception errors, reduce defects, and improve process efficiency. In this literature review, we attempt to gain an understanding of existing research on anomaly detection using open-source tools for detecting very small visual defects on thermal Integrated Heat Spreaders (IHS) lids, document challenges related to real world dataset, key learnings, and explore future opportunities in this growing field.
