Enhanced Image Patch Approximation For Lung Tissue Classification Using Feature Based Extraction

The growing size and number of the medical images necessitated the use of computers to facilitate processing and analysis. In medical world diagnostic Imaging is an invaluable and important tool for early detection of diseases. An enhanced feature descriptor that categories lung tissues in High Resolution Computed Tomography (HRCT) images used for Computer Aided Diagnosis is proposed in this paper. The images are divided into multiple Image Patches called AROI (Annotated Region of Interest).Image features like Texture, intensity and gradient are considered for feature extraction and classification. Labeling is done using a new patch-adaptive sparse approximation method. The proposed method is evaluated on a publicly available Interstitial Lung Disease (ILD) database to show the performance improvement.