An Optimized Breast Cancer Diagnosis System Using a Cuckoo Search Algorithm and Support Vector Machine Classifier
Prabukumar Manoharan, Agilandeeswari Loganathan, Arun Kumar Sangaiah · 2017
Today, breast cancer is the most important cause of cancer death for women. The early detection of breast cancer is very important to increase a patient's survival time. In this chapter, we propose a diagnosis system for early detection of breast cancer tissues from the digital mammographic breast images using a cuckoo search optimization algorithm and support vector machine (SVM) classifier. In general, the complete diagnosis process involves various stages such as preprocessing of images, segmentation of such breast cancer region from its surroundings, extracting tissues of interest and then determining the associated features that may be vital, and, finally, classifying the tissue into either benign or malignant. In our approach, for the accurate segmentation of breast cancer, the hybrid technique, namely Otsu thresholding, and morphological segmentation algorithms are used. Then the important features of the tissue of interest such as shape, statistical, texture, and invariant moments are extracted. From the above extracted features, the optimized features used for the classification of breast cancer are identified using the cuckoo search optimization algorithm. Finally, the (SVM) classifier is trained using these optimized features, which in turn helps us to classify the breast cancer of type benign or malignant. The accuracy of the proposed system is validated using Mammographic Image Analysis Society (MIAS) public database images. The overall accuracy rate of our proposed system is about 96.72%, which is high enough when compared to the existing systems.