Mammogram Diagnostics via 2-D Complex Wavelet-based Self-similarity Measures

Abstract

Breast cancer is the second leading cause of death in women in the United States. Mammography is currently the most eective method for detecting breast cancer early; however, radiological inter- pretation of mammogram images is a challenging task. Many medical images demonstrate a certain degree of self-similarity over a range of scales. This scaling can help us to describe and classify mammograms. In this work, we generalize the scale-mixing wavelet spectra to the complex wavelet domain. In this domain, we estimate Hurst parameter and image phase and use them as discriminatory descriptors to clas- sify mammographic images to benign and malignant. The proposed methodology is tested on a set of images from the University of South Florida Digital Database for Screening Mammography (DDSM). Keywords: Scaling; Complex Wavelets; Self-similarity; 2-D Wavelet Scale-Mixing Spectra.