A Deep Learning and Feature Optimization-Based Approach for Early Breast Cancer Detection

Millions of people now have better prognoses due to early and accurate diagnosis of breast cancer. Computer-aided diagnosis, an vital component of detection, eases the burden of accurate detection on medical staff. The process of diagnosing breast cancer with computer assistance involved the utilization of different machine learning like deep learning techniques. It techniques require the high level of computing complexity due to their complex processing. This study suggests the use of feature extraction and feature optimization for the diagnosis of cancer of breast. Information was extracted by using stationary wavelet transform method in the proposed methodology. The spider-monkey optimization technique was used to extract the features. A dynamic, population-based meta-heuristic for improved feature optimization of medical images is the spider-monkey optimization function. We proposed a neural network-based classifier (NNs) for cancer identification. An accuracy of identifying breast cancer is improved by the suggested neural network-based classifier. MATLAB 2018R was used to simulate the recommended approach. The method's valuation was conducted using the CBIS-DDSM cancer image dataset. the effectiveness of the suggested method compared to other breast cancer algorithms, including CNN, Ransom-RF, SVM, and NN. The suggested algorithm and the current breast cancer algorithm are both outperformed by the outcome analysis.