Disclosed is a system, method and computer readable medium for selecting a coreset of potential defects for estimating expected defects of interest. An example method includes obtaining a plurality of defects of interest (DOIs) and false alarms (FAs) from a review subset selected from a group of potential defects received from an inspection tool. The method further includes generating a representative subset of the group of potential defects. The representative subset includes potential defects selected in accordance with a distribution of the group of potential defects within an attribute space. The method further includes, upon training a classifier using data informative of the attribute values of the DOIs, the potential defects of the representative subset, and respective indications thereof as DOIs or FAs, applying the classifier to at least some of the potential defects to obtain an estimation of a number of expected DOIs in the specimen.