Breast cancer detection using active contour and classification by deep belief network

Abstract

The abnormal modifications in tissues or cells of the body and growth beyond normal grow and control is called cancer. Breast cancer is one of the kinds of cancer. The existence level of patient suffering from breast cancer can be improved through Prognosis detection of breast cancer. The cancer diagnosis and detection accuracy have upgraded with the machine learning techniques, and the improvement of expertise. A computer-aided diagnosis (CAD) system is used for the mammograms to enable initial breast Cancer recognition, analysis, and treatment. For existing CAD system the accuracy is unacceptable. This research work explores a breast CAD scheme established on feature fusion through Convolution Neural Network deep features. The outcome shows that the random forest algorithm contributes the maximum accuracy (97.51%) with the lowest error rate than CNN classifier (95.65%). The abnormality of the breast images is scrutinized through Deep Belief Network (DBN). The offered work practices active contour segmentation to distinct the abnormal image and it can be categorized by Deep Belief Network. The recommended algorithm is used to eradicate the properties of the point spread function in the method of low-dose medical CT image restoration and recovers the reconstructed image quality. The efficiency and speed of the segmentation process are increased through sparse transform. Our mentioned method is used to afford the enhanced performance, offer a good implementation of the results, evaluated the accuracy and the processing time.