Detection and Classification of Breast Cancer in Mammographic Images Using Efficient Image Segmentation Technique

Abstract

Breast cancer has become one of the major types of cancer-caused deaths among women of different countries throughout the world. One of the major problems of this type of cancer disease are quick detection or identifying of disease in early stages. In the cases of technologically lagging countries mortality rates are very high due to lack of early diagnosis technology of disease. According to the opinion of different clinical experts, today mammography is one of the most effective diagnosis technologies in medical science domain. So there is a requirement for more accurate methods which can easily diagnose any type of abnormalities in women breast without any kind of human intervention with higher accuracy rates. Segmentation is an approach that is very much required to identify the unambiguous region from the mammogram image. Intensity, texture, and shapes are extracted from the segmented mammogram image. The role of image processing is to detect cancer in human body when input data is in the form of images. For mammogram image classification, the feature extraction of an image with statistical parameter measurement is very important approach. Different types of feature extraction methods are generally used for better classification of abnormality present in mammogram. This technique will provide higher accuracy rates at a comparative higher speed. The statistical parameter includes entropy, mean, regression, correlation, skew, standard deviation. The experimental results achieved 89% accuracy, 74% specificity, and 89% sensitivity, illustrating the usefulness of the technique for identifying and classifying the cancer in mammogram images with maintaining more accuracy.