Kidney disease detection and segmentation using artificial neural network and multi-kernel k-means clustering for ultrasound images.

• To design and develop effective kidney disease detection approach. • An artificial neural network and multi-kernel k-means clustering algorithms were used. • Stone and tumor were effectively identified by medical ultra sound images. • Performance of linear + Quadratic kernel approach is analyzed with sensitivity, specificity and accuracy. • The proposed linear + Quadratic kernel approach achieves the maximum accuracy compare to other methods. The main aim of this paper is to design and develop an approach for kidney disease detection and segmentation using a combination of clustering and classification approach. Nowadays, kidney stone detection and segmentation is one of the crucial procedures in surgical and treatment planning for ultrasound images. However, at present, kidney stone segmentation in ultrasound images is mostly performed manually in clinical practice. Apart from being time-consuming, manual stone delineation is difficult and depends on the individual operator. Therefore, in this work, we proposed a kidney stone detection using artificial neural network and segmentation using multi-kernel k-means clustering algorithm. Normally, the system comprises of four modules like (i) preprocessing, (ii) feature extraction, (iii) classification and (iv) segmentation. Primarily, we eliminate the noise present in the input image using median filter. Then, we extract the important GLCM features from the image. After that, we classify the image as normal or abnormal using neural network classifier. Finally, the abnormal images are given to the segmentation stage to segment the stone and tumor part separately using multi. Kernel K-means clustering algorithm. The experimentation results show that the proposed system as linear + quadratic based segmentation achieves the maximum accuracy of 99.61%, compare with all other methods. [ABSTRACT FROM AUTHOR]