Tumor Growth Stage Prediction on Ultrasound Breast Cancer Images with Backpropagation Network

Abstract

Neural Networks are computational models for solving various complex problems. Systematic learning without user support for ultrasound-screened breast cancer images aims to predict the growth of the tumor. Even though technology is improving towards the maximum, automatic cancer prediction through systematic learning boosts disease identification. To provide quality imagery and classification of tumors, breast cancer tumor prediction is processed. Apart from a radiologist's suggestion, to give a new source for finding tumor growth, automatic learning over the system is done through a neural network. The goal of this study is to reduce errors by modifying the backpropagation network. The dataset is ready for testing, validation, and training. The neural network backpropagation is implemented with stable, bounded dependent variables to reduce the error rate of the learning parameters. The proposed methodology maintains a platform for reducing the error to a minimum for learning and classification.