Enhancing Radiologist’s Performance in Breast Cancer Screening Using PSO Algorithm with CNN Classifier

In this Research Paper, we introduce an innovative strategy aimed at improving the performance of radiologists in the crucial task of breast cancer screening. By optimizing parameters for a hybrid Particle Swarm Optimization (PSO) algorithm combined with a Convolutional Neural Network (CNN), our approach seeks to enhance the accuracy and specificity in the detection and diagnosis of breast cancer. We evaluate the efficacy of our proposed method by comparing it against a conventional approach that utilizes Deep Neural Networks (DNN), focusing on metrics such as accuracy and specificity, as well as the ability to identify both benign and malignant cases. Our findings demonstrate that the PSO-CNN method surpasses the existing DNN-based approach by achieving an accuracy rate of 95.5% compared to 92.8%. This improvement is crucial in the context of breast cancer screening where even minor enhancements in accuracy can significantly impact patient outcomes. Moreover, our method also shows an improvement in specificity reaching a rate of 90% compared to 87.4% in the DNN-based approach, thereby reducing the likelihood of false positives and minimizing unnecessary patient anxiety and follow-up procedures. The effectiveness of our approach is assessed by comparing it with a conventional system that employs Deep Neural Networks (DNN) as the primary diagnostic tool. The comparison focuses on key metrics such as accuracy and specificity, as well as the capability to detect both begin and malignant breast cancer.