AN IMAGE ANALYZING SYSTEM FOR INTERSTITIAL LUNG ABNORMALITIES IN CHEST RADIOGRAPHY - DETECTION AND CLASSIFICATION BY LAPLACIAN-GAUSSIAN FILTERING AND LINEAR OPACITY JUDGMENT

RATIONALE AND OBJECTIVES. The effectiveness of a computerized analysis system to detect and characterize interstitial lung abnormalities seen on chest radiographs was evaluated. This method included a process of four-directional Laplacian-Gaussian filtering and a process of linear opacity judgment. For quantitative analysis of interstitial opacities, the radiographic index, which is the percentage of opacity areas in a region of interest, was obtained and evaluated in the images. These opacities represented interstitial lung abnormalities. METHODS. Two regions of interest were selected in each right lung of 50 patients with normal lung parenchyma and 50 patients with diffuse interstitial lung abnormalities, confirmed with high-resolution computed tomography. These regions of interest were processed by our computerized analysis system. RESULTS. Abnormal lungs were well differentiated from normal lungs by the radiographic indices obtained from the images filtered by four-directional Laplacian-Gaussian filters and from those processed by linear opacity judgment. However, honeycomb lesions and other interstitial abnormalities (interstitial changes other than honeycombing) were differentiated from each other only by the radiographic indices obtained from the image processed by linear opacity judgment (P < .05). DISCUSSION. These results indicate that this system is useful for the detection and characterization of interstitial lung abnormalities.