Breast Cancer Cell Segmentation Using Attention-Based Deep Neural Network
Ankita Patra, Nalini Kanta Barpanda, Prabira Kumar Sethy, Ashis Das, Santi Kumari Behera, Amlan Nanda · 2023
Early diagnosis is crucial for the treatment of breast cancer, the most prevalent disease in women globally. Segmenting cancer cells from photographs of breast tissue is an important first step in the assessment of breast cancer. Here, we propose a deep neural network that uses attention to aid in the separation of breast cancer cells. The proposed technique employs an encoder-decoder architecture with an attention mechanism to zero on the most relevant parts of a picture. We tested the proposed approach on a collection of breast tissue pictures and compared the results with those obtained using the current best practices. Our experiments show that the proposed technique achieves outstanding performance, with an accuracy of 97.26 percent, sensitivity of 95.26 percent, specificity of 98.98 percent, and Jaccard index of 0.89.