DETECTING DEFECTS IN SEMICONDUCTOR SPECIMENS USING WEAK LABELING

摘要

A system for classifying a pattern of interest (POI) on a semiconductor specimen, the system comprising a processor and a memory circuit, the processor and memory circuit being used to acquire a high-resolution image of the POI and classifying the POI according to a defect-related classification Constructed to generate possible data, generating utilizing a machine learning model trained according to training samples, the training samples being a high-resolution training image captured by scanning an individual training pattern on a specimen - an individual training pattern Similar to a POI—and it includes a label associated with the image, the label derived from a low-resolution survey of an individual training pattern.