A Deep Learning-Based Approach for Medical Deepfake Detection Using YOLO Architectures

This study proposed a deep learning-based medical deepfake detection framework using multiple YOLO architectures. The rapid growth of deepfake technology has created major challenges in healthcare systems because manipulated medical images may lead to incorrect diagnosis, insurance fraud, and severe threats to patient safety.Two different datasets consisting of Knee Osteoarthritis X-ray images and Lung CT scan images were used for experimentation. Data preprocessing and augmentation techniques were applied to improve dataset quality and model robustness. Different YOLO architectures including YOLOv3, YOLOv5nu, YOLOv5su, YOLOv8n, YOLOv8s, YOLOv8m, YOLOv8l, and YOLOv8x were evaluated. Experimental results demonstrated that YOLO-based models achieved high accuracy in detecting manipulated medical images. Among all architectures, YOLOv5su achieved the best performance with a recall value of 0.997 while maintaining faster execution speed than larger YOLOv8 models. The results indicate that YOLO-based architectures provide an effective balance between computational efficiency and detection accuracy. The proposed framework can be applied in hospitals, healthcare centers, and medical imaging systems to improve diagnostic reliability and prevent malicious manipulation of medical data. Future work may include: ·       Expanding the dataset size ·       Integrating Transformer architectures ·       Real-time deployment ·       Multi-modal medical image analysis ·       Cloud-based healthcare security systems

Goodfellow, I., et al. “Generative Adversarial Nets.” Neural Information Processing Systems, 2014.Solaiyappan, S., and Wen, Y. “Machine Learning Based Medical Image Deepfake Detection: A Comparative Study.” Machine Learning with Applications, 2022.Mirsky, Y., et al. “CT-GAN: Malicious Tampering of 3 D Medical Imagery Using Deep Learning.” USENIX Security Symposium, 2019.Kim, Y. S., et al. “Development of Deepfake-Based Deep Learning Algorithm for Medical Data Manipulation Detection.” Webology, 2022.Albahli, S., and Nawaz, M. “MedNet: Medical Deepfakes Detection Using Improved Deep Learning.” Multimedia Tools and Applications, 2023.Krizhevsky, A., et al. “ImageNet Classification with Deep Convolutional Neural Networks.” NIPS, 2012.Zhu, J., et al. “Unpaired Image-to-Image Translation Using CycleGAN.” ICCV, 2017.Budhiraja, R., et al. “Medical Deepfake Detection Using Convolutional Reservoir Networks.” IEEE, 2022.Du, J. “Understanding Object Detection Based on CNN Family and YOLO.” Journal of Physics, 2018.Karaköse, M., et al. “A New Approach for Effective Medical Deepfake Detection in Medical Images.” IEEE Access, 2024.