Abnormality classification in the kidney ultrasound images using singular value decomposition features

Kidney diseases are evolving as a common chronic disease like hypertension, diabetes, and cardiovascular disease. They do not show any significant symptoms at an earlier stage. Therefore, monitoring of kidney diseases at regular interval of time is required to prevent kidney failure. This paper deals with the automatic abnormality classification in the kidney ultrasound images. The singular value decomposition (SVD) algorithm is used to extract features from ultrasound images and these features are given to the support vector machine (SVM) classifier for classification. The performance comparison of SVM is done with different classifiers along with the extracted SVD features to detect the abnormalities. The kidney classes are classified into normal and abnormal kidney with a total of 100 ultrasound images. The efficiency of the classifier is measured in terms of recall, selectivity and accuracy.