A Novel Masses Detection Algorithm Using Model-based Location and ANFIS-based Segmentation

Breast cancer is one of the most dangerous tumors for middle-aged and older women, and mammography is its most reliable early detection method. In this paper, a novel algorithm was proposed to assist the radiologists to detect the masses hidden in the mammograms. At first, two models were established to represent all kinds of masses, and according to the different type, different methods were used to detect the masses. For those masses inside the fatty tissue, iterative thresholding could be applied to locate them, while for those masses inside the denser tissue, black hole registration based on discrete wavelet transform (DWT) were utilized for their labeling. And then, filling dilation technique was used to extract the whole masses from the backgrounds, according to their labels. ANFIS, Canny edge detector and energy field method were applied to adjust the segmentation process, and the final result validated the advantages of this algorithm.