Computer-aided diagnosis of interstitial diseases: Quantitative analysis of lung textures in digital chest radiographs

The authors are developing an automated method to detect and characterize interstitial disease using physical texture measures in digital chest radiographs. Fine lung textures are isolated from the gross lung anatomy, and then objective measures related to the magnitude and coarseness (or fineness) of textures are determined from the power spectrum. The authors determined these texture measures for approximately 50 normal and 150 abnormal lungs with various patterns, as well as the ILO pneumoconiosis standard radiographs. The authors' results indicate that this method may form a basis for objective evaluation of interstitial lung disease in digital radiographs