Detection of Interstitial Lung Disease using correlation and regression methods on texture measure

A novel procedure to detect Interstitial Lung Disease (ILD) with High Resolution Computed Tomography (HRCT) images is proposed. Seven texture measures from a selected slice of HRCT lung image of fifteen ILD cases, fifteen non-ILD cases and fifteen healthy individuals were obtained and their pairwise correlation calculated. Two groups of texture measure obtained from standard clustering procedure were then used to create a discrimination procedure using simple linear regression. The texture measure Contrast and Standard Deviation of Energy (STDE) gave the highest detection rate.