Breast Cancer Detection from Mammograms Using Deep Learning and Bio-inspired Optimization Algorithm

Breast cancer is a leading factor behind cancer-caused fatalities in women all around the world. Over the past few years, the use of mammographic images for breast cancer analysis has become increasingly popular due to its noninvasive nature. In this study, we applied optimization techniques to enhance the precision of mammographic image-based breast cancer analysis. DenseNet169 and Bi-LSTM were employed for feature extraction. These are advanced machine learning models capable of automatically extracting sophisticated features from images. We also applied particle swarm optimization (PSO) for feature selection, which is a metaheuristic optimization technique that can efficiently select the most informative features for classification. To construct our model, we used one input layer, three dense layers, and one output layer. We evaluated performance of our model by measuring several metrics such as accuracy, f1-score, support, recall, precision, confusion matrix, and specificity. Outcomes of our study demonstrated that our model was able to accurately distinguish cases of breast cancer with a remarkable precision of 99.23%, suggesting that our model is highly effective in this regard. Finally, we discuss the potential for further research in this area and the implications for improving breast cancer detection and treatment. Overall, our study demonstrates that optimization techniques such as PSO, coupled with deep learning models, can significantly improve the accurateness of breast cancer analysis using mammographic images, and the hybrid approach of DenseNet169 and Bi-LSTM provides better results. This approach has the potential to aid in the early identification and handling of breast cancer, eventually saving lives.