COMPUTERIZED ANALYSIS OF INTERSTITIAL DISEASE IN CHEST RADIOGRAPHS - IMPROVEMENT OF GEOMETRIC-PATTERN FEATURE ANALYSIS

We have been developing automated computerized schemes to assist radio logists in interpreting chest radiographs for interstitial disease bas ed on texture analysis and geometric-pattern feature analysis. In this study, we attempted to improve the performance of the geometric-patte rn feature analysis, because the current classification performance wi th geometric-pattern feature analysis is considerably lower than that of texture analysis. Zn order to improve the performance in distinguis hing between normal lungs and abnormal lungs with interstitial disease , we attempted to re move rib edges in regions of interest (ROIs) by u sing an edge detection technique, and also to reduce false positives b y using feature analysis techniques. In addition, the effects of many parameters on classification performance were investigated to identify proper threshold levels, and subsequently the specificity of the geom etric-pattern feature analysis was improved from 69.5% to 86.1% at a s ensitivity of 95.0%. Using a combined rule-based method with texture a nalysis and geometric-pattern feature analysis plus the artificial neu ral network (ANN) method for classification, a high specificity of 96. 1% was obtained at a sensitivity of 95.0%. (C) 1997 American Associati on of Physicists in Medicine.