Enhancing Breast Cancer Detection through Advanced Deep Learning: An Application of YOLOv8x on Mammographic Images

This study explores the potential of advanced deep learning techniques, specifically the YOLOv8 architecture, for enhancing breast cancer detection within mammographic imaging. Utilizing the well-established MIAS and DDSM datasets, we aimed to improve detection accuracy through rigorous preprocessing and data augmentation techniques. By implementing the YOLOv8x model variant, known for its high precision, we conducted a series of experiments to assess the impact of data augmentation and combined dataset training on model performance. The experiments were structured around training the model for 60 epochs under different conditions, including original and augmented datasets, individually and combined. The performance was evaluated based on the mean Average Precision $(\mathrm{m A P})$, with the augmented combined dataset experiment yielding the highest mAP of ${9 6. 2 1 \%}$. These results demonstrate the effectiveness of combining advanced YOLO architectures with extensive data augmentation and dataset merging in improving the accuracy and reliability of breast cancer detection in mammographic images. The study underscores the significant potential of integrating AI-driven methods into diagnostic workflows, contributing to early and accurate breast cancer diagnosis.