<![CDATA[One of the new challenges resulting from the advancement of Artificial Intelligence is the creation of artificially realistic media, including the generation of fake images known as deepfakes. The generation of increasingly realistic fake images is also carried out in various fields of life including health. Detection methods are constantly being updated to address the misuse of fake medical images. This research focuses on deepfake detection using real and fake lung CT scan datasets. The YOLOv11 model is tested incrementally to be able to detect both types of images. Two manipulation methods CT-GAN and stable diffusion (SD) are used to test the performance of the YOLOv11 model. The results showed that the YOLOv11 model tested using stable diffusion artificial image manipulation achieved 100% accuracy, precision, recall and f1-score. Meanwhile, CT-GAN image manipulation has problems in producing a perfect model in distinguishing real and fake lung cancer CT scan images. With further improvements and enhancements, YOLOv11 fine tuning results can be one of the options for deepfake medical image models that are relatively lightweight, fast, and accurate.]]>
Copyrights © 2026
