Detection of Breast Cancer from Histopathological Images
Agna Shaju P., Subhija E.N. · 2022
Breast cancer is the second leading cause of death after lung cancer, and breast cancer accounts for around 30 per cent of newly diagnosed cancer cases. Since manually detecting a cancer cell is time-consuming and prone to error, computer-assisted processes are used to get better results than manual pathological detection systems. Breast cancer is detected via Mammography, MRI scans, CT scans, ultrasounds, and nuclear imaging, but none of these guarantee a 100 per cent accurate diagnosis. For tumor detection, a histopathology test is performed to detect invasive cancer cells using HE stained tissues. However, intra-observer variation, numerous appearances of cancer cells or tissues, and the same hyperchromatic highlighted cells make identification challenging. Deep learning algorithms can be used to solve these issues. Deep learning uses a Convolutional Neural Network (CNN) to extract information before classifying them with a fully connected network. Deep learning is widely used in the field of medical imaging, and it does not require prior knowledge in a related discipline. The purpose of this automatic retrieval approach appears to have practical utility in the field of imaging for diagnostic imaging professionals in determining whether a tumour is benign or malignant in nature. The objectives of the project include the following: 1. Learn about different deep learning methods used in analysis of histopathological images for medical diagnosis. 2. Detect breast cancer from histopathological images using deep learning. 3. Classify and identify different types of cancerous cell using the algorithm. 4. Comparing the method with other existing methods and analyse the performance.