Breast Cancer Image Classification Using Deep Convolutional Neural Networks

Abstract

In recent years, convolutional neural network algorithms have made remarkable progress in the classification of medical images such as the classification of breast cancer tumors. Models of deep convolutional neural networks have obtained a higher accuracy rate in medical image recognition. The fine-tuning of images data and parameters are the main task of adapting a pre-trained convolution model in order to improve the classification accuracy. This paper aims to present a model for the use of deep neural networks, specifically convolutional neural network model AlexNet, for breast cancer classification. The model will be used to diagnose breast cancer using histopathological BreakHis images data set. Modifications of parameters and data are applied to increase the model ability for recognizing and classifying the input image and determine whether the image belongs to a benign or malignant tumor. It has been noticed that the training frequency and balanced training data greatly improve classification rate accuracy up to 96%. Our mission is that to achieve a higher accuracy rate than the obtained if repeated improvement of fine-tuning parameters and weights are adopted according to more accurate techniques.