A new ANN-based detection algorithm of the masses in digital mammograms

Breast cancer has become one of the most dangerous carcinomas for middle-aged and older women in China recently. Mammography is its most reliable detection method in the clinic, and computer-aided diagnosis (CAD) could assist the radiologists in reading the mammograms. In this paper, a new algorithm based on two ANNs (artificial neural networks), was proposed to detect the masses automatically. It firstly built up two mass models to represent the masses with different backgrounds and features, and used different detection methods on different type of masses: for those masses inside the fatty tissue, iterative threshlding was applied to locate them; for those masses in the denser tissue, black hole registration based on discrete wavelet transform (DWT) were used instead. Then, filling dilation was used to extract the whole masses from the backgrounds, which was adjusted adaptively by ANFIS (adaptive-network-based fuzzy inference system). At last, the segmented suspicious masses were filtrated with a MLP (multilayer perceptrons) classifier. With these two ANNs, the detection process were well adjusted and improved, and the final diagnosis result showed that the CAD scheme could simultaneously achieve comparatively high detection precision and low false positive rate, even when the special masses were dealt with.