Automated Mammogram-Based Breast Cancer Detection with Deep Learning and Advanced Image Enhancement

• DOI: 10.52783/jisem.v10i45s.9013
• Publication Date: 2025-05-14T11:27:36Z
• Authors: Hazheen Sarbast ...

Breast cancer is a significant worldwide health issue, and early detection key to enhancing survival. Although mammography the accepted screening method, human interpretation susceptible errors, resulting in misdiagnosis. Convolutional neural networks (CNNs), particular, have shown promise deep learning for automating breast detection, increasing accuracy, reducing variability. In this research, model automatically classifying from mammograms proposed evaluated. The suggested model's performance on CBIS-DDSM dataset compared transfer using pre-trained models like MobileNetV2, DenseNet121, EfficientNetV2L terms of classification accuracy generalizability. Moreover, study investigates how data augmentation preprocessing affect models. Accuracy, sensitivity, specificity computational efficiency were used measure performance. technique achieved highest performance, with 98.21% 99.04% Sensitivity 97.33% Specificity 80%-10%-10% split 99.02% 85%-5%-10% split. Affirming effectiveness 99.24% 98.80% detection. This systematically contrasts CNN mammogram classification, optimizes methods, evaluates efficiency. It fills research gap by balancing against tractability proves higher diagnostic potential AI-augmented mammography. verifies that drastically enhance based mammograms, being most balanced between