Breast Cancer Detection And Classification Using Artificial Neural Network

Abstract

The main objective of the paper is to analyze the performance of the Artificial Neural Network model in medical image diagnosis. This paper helps beginners understand basic architecture and the functionality of ANN who want to start learning deep learning. The high-level neural network API is created using Keras on top of low-level APIs like Tensorflow. The high-level architecture of ANN has many built-in libraries and functions to do image processing tasks efficiently and effectively, and it is considered an excellent tool for diagnosing breast mammogram images. This paper uses Keras to do perfect abstraction of creating deep learning operations. The conventional methods used in the earlier research have not provided efficiency in automatic, less complexity, and improved accuracy. Also, they used multiple sub-modules to perform separate individual functions to classify cancer, which increases the complexity in terms of time, coding, and cost. Thus, this paper aims to implement an Artificial Neural Network for learning, analyzing, and classifying breast cancer from WDBC and Mammogram data. The experimental results proved that the ANN outperformed other conventional methods and obtained 100% and 98.56% accuracy on WDBC and mammogram images, respectively.