A novel neural network based automated system for diagnosis of breast cancer from real time biopsy slides

Breast cancer is one of fatal disease in women, which is better curable if detected at an early stage. This paper presents a neural network based diagnosis system for breast cancer. Neural network system has the ability to be trained by large data and hidden information or features in the samples. Thus exhaustive case studies of a specialized doctor can be used to train a neural network which results in an efficient decision making tool in field of cancer diagnosis. Artificial neural network (ANN) based diagnostic system for breast cancer is developed using two stages. Firstly, the well-known Wisconsin Breast Cancer Database (WDBC) is used to develop the diagnostic system. Supervised training of neural network using back propagation algorithm is done. Two variants of back propagation algorithm are investigated with three and four layers of neural networks. With efficient neural network developed, the tested method is applied for biopsy slide images of breast tissues. Preprocessing of images is done to extract required key features using image processing algorithms using Matlab Simulink Thus a data set is created using the case studies of breast cancer cases. This work demonstrated that neural network can be an efficient decision making tool in field of cancer diagnosis using biopsy slides.