APPLICATION OF WAVELET TRANSFORM AND NEURAL NETWORKS TO DETECT AND DIAGNOSE MICROCALCIFICATIONS IN MAMMOGRAMS

Abstract

Breast cancer accounts for the most cancer diagnoses and the second most cancer deaths of women. Digital mammography is one of the most suitable methods for early detection of breast cancer. However, it is very difficult to distinguish benign and malignant microcalcifications (MCs). There are many deaths caused by late detection or misdiagnosis. An intelligent computer-aided diagnosis system (CAD) can provide a second opinion to the radiologists. Given that the MCs correspond to high frequency components of the image spectrum, detection of MCs is achieved by decomposing the mammograms into different frequency subbands, suppressing the low frequency subband, and, finally, reconstructing the mammogram from the high frequency subbands. A combination of 3 features (variance, entropy and standard deviation) computed by discrete wavelet transform are used as inputs to a simple neural network consiting of one hidden layer with 5 nodes. The system performs well with the accuracy of about 91% on data images. The result shows the advantage of wavelet transform associated with neural network in CAD system for mammography.