Classification of Breast Cancer using Ridgelet based Feature Extraction

Breast cancer is the most common cancer all over the world. Ultrasound imaging is an effective tool in the detection of breast cancer because of its non ionizing radiation and low cost. This paper presents the automatic classification of breast cancer by ridgelet feature extraction method. Speckle noises in the ultrasound images are reduced by mean, Weiner and filtering techniques and region of interest is cropped from the filtered image. Statistical texture features are extracted by applying ridgelet transform. Using the feature set abnormalities are classified SVM classifier. The Performance of the feature selection methods best feature selection method is evaluated based on the results.