Breast cancer detection in pathological imaging using deep learning methods
Bharati Ainapure, Reshma Nitin Pise · Institution of Engineering and Technology eBooks · 2023
Breast cancer is a disease that affects many women globally and is associated with a significant fatality rate. Worldwide nearly 12% of women are affected by breast cancer and the number is still increasing. To enhance breast cancer detection and patient survival rates, early and accurate identification of the disease is essential. Hence, there is a need for diagnostic models based on medical imaging which will help medical practitioners in diagnosing and treating the patients with minimum error and greater accuracy. Artificial intelligence (AI) and machine learning (ML) techniques can be used in medical imagining detecting complex relationship between different data elements so that disease detection and prognosis can be made easy. Deep learning (DL) techniques can be used to detect the most influencing features from images so that most serious diseases such as breast cancer can be treated in-time. In this chapter, a systematic evaluation of prior work based on breast cancer identification and prognosis using images like mammography, MRI, etc. along with DL and ML is carried out. Based on this the work proposes a design of three advanced DL methods: convolution neural network, ResNet, and U-Net to classify breast cancer mammography images. The models were trained to identify histopathology images into to two classes: malignant and benign. The method is implemented in two steps. The first step comprises of data selection and pre-processing and, in the second step, the implementation of three networks and performance measurement of each network. The results indicate U-net model outperformed by achieving 97.12% of accuracy compared to other networks.