NASNet-Driven Classification Framework for Accurate and Early Breast Cancer Diagnosis

Breast cancer detection is crucial in minimizing world mortality rates and enhancing patient prognosis by means of early diagnosis and prompt treatment. The creation of dependable, accurate, and computerized detection systems is of primary concern in contemporary healthcare due to the fact that it greatly helps clinicians detect malignant lesions at an early stage. But current deep learning and computer-aided diagnosis (CAD) systems have limitations in the form of low generalization, hand-designed architecture, high false positive rates, and poor performance across various mammographic views. Towards overcoming the stated challenges, the present paper introduces a new NASNet-based breast cancer detection and classification system as part of the existing cutting-edge deep learning models possessing the ability to generate optimized network structures for image-related processes on an automatic basis. The new system is all set for testing of mammography images with higher precision, sensitivity, and specificity. The innovation is interested in using NASNet in the health industry with dynamic flexibility and higher diagnostic accuracy performance. Experimental outcomes indicate that the proposed NASNet model works greatly, with 96% accuracy rate, $\mathbf{9 6 \%}$ precision rate, $\mathbf{9 6 \%}$ recall rate, $\mathbf{9 6 \%}$ F1-score rate, 97% AUC, and 96% specificity, surpassing a number of state-of-the-art deep learning methods.