SePArate: Segmenting Patterns from Defects in Wafer Manufacturing Using Weak Supervision
2026-08-31 • Computer Vision and Pattern Recognition
Computer Vision and Pattern RecognitionArtificial Intelligence
AI summaryⓘ
The authors developed SePArate, a new method to help find and separate defects on semiconductor wafers using only basic image labels instead of detailed pixel-by-pixel markings. Their method trains in three steps: first learning general features, then learning to identify locations of defects, and finally training on fake mixed-defect images to improve accuracy. Their tests show it performs better than existing methods at pinpointing defect areas. This approach could make defect analysis more scalable and precise.
semiconductor manufacturingdefect analysiswafer defect segmentationweakly supervised learningencoder pretrainingknowledge transfersynthetic datapixel-level segmentationimage-level annotation
Authors
Dain Kwon, Changmin Shin, Sunjong Park, Kanghyun Choi, Hyeyoon Lee, Jaewon Jang, Minseok Choi, Jinho Lee
Abstract
In semiconductor manufacturing, defect analysis is essential, but manual inspection cannot scale. However, existing automated inspection methods remain insufficient for root-cause analysis and process optimization. To this end, we present SePArate, a weakly supervised wafer defect segmentation method. SePArate enables pixel-level separation of patterns by leveraging only image-level annotations. It consists of a three-phase training: encoder pretraining, knowledge transfer to learn spatial cues, and training on synthetic mixed-defect data for accurate segmentation. Experiments demonstrate that SePArate outperforms the baselines.