Breast Cancer Classification by Implementation of Deep-Learning with Dataset Analysis

Abstract

Cancer is a fatal disease recognized and researched about, around the globe. Researchers and scientists have been investing their time and imparting their expertise, and knowledge for the advancements of traditional methods and treatments to tackle it. Recent surveys reveal that the mortality rate among the female populous, over the world, is also one of the results of breast cancer. The definition of breast cancer can be described as an uncontrolled aggressive growth of old cells which thereby aid the formation of a pernicious mass in the tissue of a breast. Gradually, this may result in the formation of a tumor of malignant nature. Deep learning, considered a sub-field of Machine Learning, enables experts to analyze, model, and study complicated or rather complex scientific data over a comprehensive list of medical applications. This study aims to create a user-friendly, adept system to perform the classification of breast tumors of malignant or benign nature. The proposed system is divided into two halves or stages. The initial stage is the pre-processing and analysis of the acquired dataset which also involves training of the neural network. The next and final stage is the classification of breast tumors by utilizing the created model and loading it onto an API through which users can upload tissue images and check what type of breast cancer the tissue contains. This would eliminate the time spent on studying every particular data using traditional clinical methods. This project would help support the radiologists in training, research, and diagnostic aspects and overall support the entire process of cancer diagnosis and treatment.