Breast Cancer Tissue Detection System Based on K-Mean Clustering and Deep Neural Network

The aim of the research is to utilize deep learning techniques to support radiologists in enhancing the effectiveness and precision of breast cancer diagnosis. This paper employs a deep neural network (DNN) classifier approach to accurately categorize breast cancer in the MIAS dataset. Firstly, we start by pre-processing the mammography images to remove the digitization noise using a Wiener filter. Afterwards, the k-mean clustering approach is used to identify the Region of Interest (ROI), from which 24 features are extracted using the color and texture feature technique. The next phase based on training a Deep Neural Network (DNN) classifier using the significant features extracted from the training set. The test set is classified the breast cancer tis-sue by using the trained DNN classifier. We used the MAIS dataset for experimental purposes. The experimental results indicate that the average classification accuracy of the proposed deep neural network (DNN) classifier is 94%. The system achieved better performance in comparison to other classifiers, achieving an accuracy of 0.94, recall of 0.87, specificity of 0.94, precision of 0.90, and F-measure of 0.88. After analyzing these measures, it is clear that the proposed DNN classifier performs better than current state-of-the-art methods.