1. JISHNU ROYCHOWDHURY - Bachelor’s degree program in electronics and communication engineering in Narula Institute of
Technology, India.
2. TATHAGATA SARKAR - Bachelor’s degree program in electronics and communication engineering in Narula Institute of
Technology, India.
3. KAUSHIK SARKAR - Associated with Narula Institute of Technology, Dept. of Electronics & Communication Engineering.
4. SURAJIT BARI - Assistant Professor, in the Department of Electronics and Communication Engineering of Narula Institute
of Technology, Agarpara, Kolkata.
5. PRANAB HAZRA - Associated with Narula Institute of Technology, Dept. of Electronics & Communication Engineering.
6. Dr. ANILESH DEY - Associate Professor of Electronics and Communication Engineering at Narula Institute of Technology,
Agarpara, Kolkata.
Breast cancer is one of the top causes of death of women globally; however it is highly treatable if detected at an early stage. The traditional approach for breast screening is x-ray mammography, which is known to be difficult for detecting cancer tumors early. The complex breast structure caused by the process of compression during imaging makes it harder to detect minor size disorders. Furthermore, inter- and intravariation of breast tissues makes it challenging to attain high diagnostic accuracy using hand-crafted characteristics. In this paper, we propose a new method with the ease of deep learning, B-Cir U-Net inspired from the U-Net architecture, for the effective and early detection, localization and classification of breast cancer. The results show a high accuracy rate along with a high rate of sensitivity and specificity, proving that the proposed methodology might be effective in medical settings.
Augmentation, Breast Cancer, Convolution Neural Networks (CNN), Deep Leaning, Image processing, Mammography, Separable Convolution, U-Net architecture.