Analysis of false-positive microcalcification clusters identified by a mammographic computer-aided detection scheme

摘要: The accuracy of computer-aided detection (CAD) schemes involves a tradeoff between high sensitivity and low false-positive rate. In an on-going study, we are analyzing our CAD scheme for the clustered microcalcifications in digital mammograms to determine causes false-negative clusters. Two different limitations that lead false-negatives false-positives have been identified. first limitation is imposed by quality mammogram, whereas second consequence similarities radiographic features true false this paper, examine effects image quality, particularly noise, on performance scheme. Preliminary results indicate limited anatomic noise x-ray quantum noise. Almost all positives detected clinical images caused combination these two forms