Statistical texture feature-based retrieval and performance evaluation of CT brain images

Advances in medical imaging technology and computer science have greatly enhanced interpretation of medical images, and contributed to early diagnosis. The development of Computer Aided Diagnosis (CAD) systems for assisting the physicians for making better decisions has attracted lot of interest in the recent past. CBIR-based CAD systems that use CBIR to search for clinically relevant and visually similar images (regions) depicting suspicious lesions have been attracting research interest. Content-based image retrieval (CBIR) of medical images is an important alternate and complement to traditional text-based retrieval using keywords. We have implemented CBIR system based on the effective use of texture information within the images. Two different statistical methods implemented are: Haralick Statistical Co-occurrence Matrix and Tamura method. To speed up the search process, selected features are extracted and indexed using hash structure. Euclidean distance measure is used for similarity measurement. Both the methods are compared based on precision and recall. Tamura features are found to provide better retrieval results for CT scan brain images.