Evaluation of Tissue Characteristics of Kidney for Diagnosis and Classification Using First Order Statistics and RTS Invariants

Analysis of abdominal ultrasound kidney images is made to evaluate the tissue characteristic for implementing unbiased diagnosis procedure and to classify important kidney orders. A set of features are estimated by quantifying spatial gray level distribution using first order statistics and algebraic moment invariants. The images are acquired from male and female subjects of age 45plusmn15 years and are pre-processed prior to feature extraction. The results obtained show that six features out of fifteen are highly significant (p<0.0005) in discriminating inter class I and III, compared to inter class II which shows significant performance (p<0.002). The second order polynomial regression analysis is also performed to measure the data stability against kidney area. The study reveals that derived features are efficient for tissue characterization of kidney which further enhance the objective diagnosis and help for designing computer-aided diagnosis (CAD) system