A Framework to Figure Out Breast Cancer Cells Using Ultrasound Images
Babe Sultana, Farzana Rahman, Hasib Ahmed Nesar, T. M. Shahriar Sazzad · 2020
Cancer shows the unregulated existence of a cluster of cells in a specific area of the body parts. Research says that among all cancers, breast cancer is the most happening cancer in this world that brings a dead percentage high every year. It is a general issue not only for women but also for men, so, early detection of breast cancer cells is highly appreciated work in both computer science and medical science. To diagnose breast cancers, electronic modalities are incorporated. Ultrasound scanning is regarded among the most generally used digital methods for breast cancer identification which is combined with less cost and protection. In this research paper, we have made a proposal of a computational framework to detect cancer cells using gray-scale ultrasound images of the breast. We have used some image processing techniques for feature extraction from potential cancerous regions and classified them using SVM classifiers. The enhancement process was initially used to optimize gradients on a gray-scale. The median filter has been used to promote better segmentation to prevent unwanted peak noise. Segmentation based on thresholds has used since the sample images used in this research are gray. We have implemented our strategy in MATLAB and tested with the SVM classifier to verify, this proposed strategy achieved 95.37% accuracy that outperforms the satisfactory level by comparing it with existing works and this accuracy is accepted by pathology experts.