Domain Adaptation for Image Classification of Defects in Semiconductor Manufacturing

In the semiconductor sector, due to high demand but also strong and increasing competition, time to market and quality are key factors in securing significant market share in various application areas. Thanks to the success of deep learning methods in recent years in the computer vision domain, Industry 4.0 and 5.0 applications, such as defect classification, have achieved remarkable success. In particular, Domain Adaptation (DA) has proven highly effective since it focuses on using the knowledge learned on a (source) domain to adapt and perform effectively on a different but related (target) domain. By improving robustness and scalability, DA minimizes the need for extensive manual re-labeling or retraining of models. This not only reduces computational and resource costs but also allows human experts to focus on high-value tasks. Therefore, we tested the efficacy of DA techniques in semi-supervised and unsupervised settings within the context of the semiconductor field. Moreover, we propose the DBACS approach, a CycleGAN-inspired model enhanced with additional loss terms to improve performance. All the approaches are studied and validated on real-world Electron Microscope images, considering the unsupervised and semi-supervised settings, proving the usefulness of our method in advancing DA techniques for the semiconductor field. Note to Practitioners—Computer Vision-based inspection systems are fundamental to ensuring quality in many productive systems. Reliable image labeling is unfortunately very hard to automatize in complex manufacturing scenarios and often still relies on human experts, especially for complex and varying image content. Hence, limited or absent labels pose a significant challenge. In this work, a Deep Learning-based approach is proposed for image classification that can exploit heterogeneous data. In the context of fault detection, the proposed approach allows for enhanced reliability and stability in the classification, which is of crucial importance for dependable automated decision-making in manufacturing. The approach has been validated on real-world data associated with scanning electron microscope images collected in a productive semiconductor manufacturing environment.