Method and system for defect classification

A method for determining the sufficiency of defect data for classification comprising:

• acquiring a set of defect data from a specimen, the defect data including imagery data associated with a plurality of defects including a plurality of defect types;
• receiving a signal from a user interface device indicative of a manual classification of the plurality of defects;
• distributing defect data of at least a first defect type into N groups of data;
• identifying a group of the N groups of data as containing test data;
• identifying N-1 groups of data of the distributed defect data not identified as containing test data as containing training data;
• for at least a first group of the N groups, incrementally generating a series of classifiers based on the training defect data contained in the N-1 groups of data, wherein each classifier is generated with an incremented percentage of at least the first defect type contained within the training defect data of the N-1 groups of data;
• determining an accuracy value for each of the series of classifiers for at least the first defect type by applying each of the series of classifiers to the test data not contained in the N-1 groups of the distributed defect data;
• and
• generating a defect data sufficiency score, for at least the first defect type, based on a generated accuracy score for at least the first group of N groups and at least one additional generated accuracy score for at least one additional group of the N groups.