SVM-based characterisation of liver cirrhosis by singular value decomposition of GLCM matrix
International Journal of Artificial Intelligence and Soft Computing
| Hi-index | 0.00 |
Ultrasonography is one of the safest methods used in imaging human organs. The visual interpretation of B-scan images principally depends on the ability of the clinician to observe certain textural characteristics. However, a visual criterion of diagnosing liver diseases primarily depends on the clinical experience of physicians and it is extremely subjective. Therefore, a method that can provides a possibility of abnormalities is valuable for physicians. Since the sets of pattern misclassified by the different classifiers designs potentially offered complementary information about the pattern to be classified this could be harnessed to improve the performance of the selected classifiers. In this study, we intend to integrate the real numerical outputs of different classifiers in order to assign a better quantitative measure of certainty about an object in question. And, a quantitative characterization about suspicious disease can be provided to a physician that can decide whether need further examination. Keywords: Multiresolution fractal feature vector, Bayes classifier, Fuzzy K-NN classifier, BPNN, MPNN, classifier combination, fuzzy integral.