Semiconductor defect detection using Machine Learning
Dr. Goddard and Dr. Schwing have developed a machine learning technique which requires only few training images to detect and classify nano-scale anomalies in semiconductors and in noisy optical images. It is able to automate defect inspection in a 9nm semiconductor wafer with high accuracy and at a high speed in fabrication laboratories using visible light microscopy. This method is faster than atomic force microscopy and scanning electron microscopy, and is more sensitive than current optical microscopy. It can also be used for other applications such as dimension measurements and disease sample measurements.