OPTIMIZATION OF UNKNOWN DEFECT REJECTION FOR AUTOMATIC DEFECT CLASSIFICATION, CAPABLE OF ANALYZING MANUFACTURING DEFECTS

摘要

PURPOSE: Optimization of unknown defect rejection for automatic defect classification is provided to perform automated classification of defects on the basis of test data.;CONSTITUTION: A computer system (800) receives confidence threshold for automatic classification of at least one defect related with class. The computer system selects the value of a parameter related with the confidence threshold. The computer system receives test data about a number of defects detected in one or more samples under test. The computer system classifies the defects about the class automatically, by using the selected value of the parameter and a kernel function.;COPYRIGHT KIPO 2013;[Reference numerals] (200) Prediction system; (802) Processor; (804) Main memory; (806) Static memory; (808) Bus; (810) Video display; (812) Zero-number input device; (814) Cursor control device; (818) Data storage device; (822) Network interface device; (824) Machine-legible storage media; (826) Commands; (AA) Signal generating device; (BB) Network