Extraction and dimensionality reduction of features for Renal Calculi detection and artifact differentiation from segmented ultrasound kidney images

Ultrasound Medical images are populated with many unwanted features and it is important to extract only the relevant information from it to reduce the complexity. In this work, RCD-AD (Renal Calculi Detection-Artifact Differentiation), Fourteen features are extracted and these extracted attributes are analyzed to give some convention for the problem which is essential for knowledge extraction. GLCM (Gray Level Co Occurrence Matrix) is used for the extraction of features from the segmented ROI (Renal Calculi/Artifact) which serves as a statistical tool to pull out the Second order textural features. Clustering based feature discretization using k-means is performed on these different classes of attributes which are extracted from the system. As the number of features increase, the amount of computation also increases. A good quality feature depends on the statistical measures such as feature variance, Inter class variation and Intra class variation. Principal Component Analysis is used as a Feature reduction technique in RCD-AD medical images.