An accurate method of breast cancer detection from ultra sound images using probabilistic fuzzy clustering algorithm

Breast cancer is the 2
^nd
leading cancer diagnosed among world. The available methods like Mammogram, MRI, CT, micro wave imaging and so on can detect both cysts and tumors in breast but possesses certain demerits. Ultra Sound is considered as an efficient way of breast cancer detection. This paper forward a method to classify benign and malignant tumor based on its boundary. Adaptive histogram equalization is used to enhance contrast. Hybrid filter is used to eliminate speckle noise. Probabilistic fuzzy clustering algorithm is employed to segment the ultra sound images. Morphological operations are done to extract specific regions. Comparison of hybrid filtered and homomorphic filtered images are done to find out the best speckle noise reduction filter.