Research On Microcalcification Enhancement And Detection

The breast cancer is one of the most common malignancy of women, which seriouslyaffect women’s health even life-threatening. Micro-calcifications are the important indicatorof the breast cancer in early stage, at the same time, is the effective diagnostic tool. However,the mammographic images usually suffer from poorly defined micro-calcification features,breast tissue and muscles. So, the Computer-aided detection is being developed to help theradiologists avoid overlooking a cancer.A series of key techniques are deeply described in Computer-Aided-Detection aboutdetecting and classifying the microcalcifications in the paper. A method is proposed to locatethe microcalcifications by enhancement, extracting, classification. The dynamic rangecompression preserving local image contrast is carry out in the processing o f enhancement;segmentation of the microcalcifications region is in the method of2-D wavelet, non-linearand statistics features;the gray level statistics features,gray level dependence matrix features,Fourier transform features and wavelet transform features are mainly proposed in theextracting features;the SVM classifies the microcalcifications region with accuracy. Based onmathematical morphology,1LOG-2LOG operator and corner detection algorithm, theaccurate location and shape of breast microcalcifications position is realized anddiscriminated finally. The result by comparing with the result of subjective detected by doctorshows that the algorithm of the breast microcalcification detection in this paper is effectiveand practical.