Severity Analysis of Mitral Regurgitation Using Discrete Wavelet Transform

This paper exhibits a computer-aided diagnosis system for the severity analysis of mitral regurgitation (MR) and assesses the discriminatory capability of Daubechies wavelet-based texture modeling. The Daubechies wavelet family has been utilized for the image decomposition because of its approximate shift invariance property. Seven statistical texture features have been utilized after the decomposition of the image up to four levels and after that concatenated. A supervised classifier, support vector machine (SVM) has been utilized with 10-fold cross-validation approach. The highest classification accuracies are 99.12 +/- 0.44 utilizing db2, 99.70 +/- 0.29 utilizing db4, and 97.68 +/- 1.04 utilizing db4 wavelet in A2C, A4C, and PLAX respectively. The exploratory outcomes show that the proposed algorithm is effective and db4 beat among the Daubechies wavelet family considered amid this audit for precise severity investigation of the MR images.