Feature extraction and analysis of renal abnormalities using fuzzy clustering segmentation and SIFT method

Ultrasound Imaging is one of the most widely used technique to provide information about renal diseases in kidney such as cyst, tumor and calculi. This paper aims to extract features from the different renal abnormalities to discriminate between the normal and abnormal conditions. Two filters, median and Wiener filter are used to remove the speckle noise in US (ultrasound) images. A picture quality performance technique is implemented to identify the quality of the images. Peak to signal noise ratio (PSNR) and Mean squared error (MSR) is used to verify the enhanced images. The preprocessed images are then segmented using FCM (fuzzy c-means clustering technique) which yields better results to find Region of interest (ROI). The statistical features along with scale invariant feature transform (SIFT) features and Texture features are extracted and analyzed. It is found that features like Energy, Variance and kurtosis are seen to be higher in the normal kidney images than the renal abnormalities. The features can be used to discriminate between normal and abnormal renal conditions. The developed system is expected to provide support for the medical practitioners for decision making to provide an enhanced health care.