Enhanced DeepFake Image Detection via Swin-B Transformer with Visual Attention Analysis. JETIA. 2026;12(58):304-314. doi:10.5935/jetia.v12i58.3045