Multiscale and Multilevel Wavelet Analysis of Mammogram Using Complex Neural Network

摘要

Mammography is an effective tool for early detection of breast cancer. The various abnormalities such as Microcalcification, Clusters, Masses, Spiculated lesions, Asymmetry and Architectural distortions are strong markers of breast cancer. Efficient diagnoses of these abnormalities from mammograms rely heavily on the kind of features extracted and the selection of classifier. In this paper, a novel methodology for microcalcification detection using multilevel wavelet analysis and Phase Encoded Complex Extreme Learning Machine is proposed. Generally, complex neural network operates only on complex features for classification. However, PECELM enables transforming the real-valued features to the complex domain to exploit the orthogonal decision boundaries of complex-valued classifiers for solving real-valued classification problems. This proposed methodology based on multiscale and multilevel Wavelet analysis on complex domain achieves an average efficiency of 95.41% and a maximum efficiency of 100%.