QUANTITATIVE-ANALYSIS OF GEOMETRIC-PATTERN FEATURES OF INTERSTITIAL INFILTRATES IN DIGITAL CHEST RADIOGRAPHS - PRELIMINARY-RESULTS

We are developing a computerized method for detection and characteriza tion of interstitial diseases based on a quantitative analysis of geom etric features of various infiltrate patterns in digital chest radiogr aphs. In our approach, regions of interest (ROIs) with 128 x 128 matri x size (22.4 mm x 22.4 mm) are automatically selected, covering periph eral lung regions. Next, nodular and linear opacities, which are the b asic components of interstitial infiltrates, are identified from two p rocessed images obtained by use of a multiple-level thresholding techn ique and a line enhancement filter, respectively. Finally, the total a rea of nodular opacities and the total length of linear opacities in e ach ROI are determined as measures of geometric pattern features. We h ave applied this computer analysis to 72 ROIs with normal and abnormal patterns that were classified in advance by six chest radiologists. P reliminary results indicate that the distribution of measures of geome tric-pattern features correlate well with radiologists' classification . These early results are encouraging, and further evaluation hopes to establish that this computerized method might prove useful to radiolo gists in their assessment of interstitial diseases. Copyright (C) 1996 by W.B. Saunders Company