AIBIZ, an AI (artificial intelligence)-based semiconductor manufacturing solutions company, announced on the 24th that it has completed a patent registration related to semiconductor process anomaly diagnostic technology. This patent concerns a method and device for automatically classifying wafer map patterns based on multiple defect points. It is designed so that an artificial intelligence model can automatically recognize and classify the various defect patterns that occur in semiconductor manufacturing processes. The core of the technology is to binarize wafer map images collected from multiple process tools, cluster patterns based on the location and shape information of defect points, and distinguish defect types with a deep-learning-based classification model. Unlike the existing method of classifying by eye depending on operators' experience, it can automatically recognize even previously unknown new defect patterns, increasing the speed and accuracy of process response. In addition, by applying defect point clustering and a training data adjustment technique that takes data imbalance into account, it addressed the problem of reduced classification accuracy caused by variation among defect patterns. As a result, stable anomaly diagnosis and process root cause analysis are expected to become possible at manufacturing sites, AIBIZ said. AIBIZ plans to apply this patented technology in stages to its AI solution for semiconductor manufacturing process diagnostics, 'DutchBoy S', to advance its wafer defect pattern analysis and process anomaly root-cause tracing functions. It also plans to expand the scope of application to process equipment condition diagnosis and analysis of the causes of yield decline. AIBIZ CEO Ha Seung-jae said, "We have completed our 24th patent registration related to semiconductor process anomaly diagnostic technology," and "This patent is a core technology that enables AI to learn and classify, on its own, the complex defect patterns that occur in semiconductor manufacturing processes." He added, "By continuously securing patents, we will raise technological barriers to entry in the semiconductor manufacturing AI field and strengthen our global competitiveness."