Data-Efficient Transformer Architectures for Image-Level Facial Forgery Detection: A Comparative Evaluation of ViT and DeiT | ||
| Journal of Information Technology Management | ||
| دوره 18، شماره 1، 2026، صفحه 17-33 اصل مقاله (1.26 M) | ||
| نوع مقاله: Research Paper | ||
| شناسه دیجیتال (DOI): 10.22059/jitm.2026.106252 | ||
| نویسندگان | ||
| Akshatha G* 1؛ Kempanna M2؛ Preethi Kolluru Ramanaiah3؛ Bandi Doss4؛ Ramani S5؛ Kirubakaran Rangasamy6 | ||
| 1Assistant Professor, Department of Computer Science& Engineering, B.M.S College of Engineering, Affiliated to Visvesvaraya Technological University, Belagavi, India. | ||
| 2Associate Professor, Department of Artificial Intelligence and Machine Learning, Bangalore Institute of Technology, Visvesvaraya Technological University, Bangalore, India. | ||
| 3Cloud Architect, Lead of AI initiative Program, Ernst & Young LLP, New York, USA. | ||
| 4Department of ECE, CMR Technical Campus, Hyderabad, Telangana, India. | ||
| 5Professor, ECE Department, Sreenidhi Institute of Science and Technology, India. | ||
| 6Assistant Professor, Department of Biotechnology, Vinayaka Mission`s Kirupananda Variyar Engineering College, Salem (Vinayaka Mission`s Research Foundation). India. | ||
| چکیده | ||
| The rapid development of deepfake technologies has increased the demand for a credible and inter-pretable system for facial forgery detection. This study compares two transformer-based architec-tures—Vision Transformer (ViT) and Distilled Data-Efficient Image Transformer (DeiT)—for de-tecting real and manipulated facial images. The study aims to measure performance in terms of de-tection as well as interpretability and to address the weaknesses of traditional convolutional models. Data augmentation was applied, and a balanced dataset containing 8,000 real and fake images was constructed; both models were then fine-tuned under the same training environment. The explanatory ability of the models was incorporated using LIME. Experimental findings indicate that both models perform well, with DeiT being slightly more accurate at 94.62% than ViT at 93.6%, alongside faster convergence rates and less overfitting. Visualization of the focus on important facial areas confirms that the models reliably register synthetic artifacts. Although promising, generalization across dif-ferent datasets and enhancement of real-time performance remain challenges. Overall, the results validate transformer architectures—especially DeiT—as powerful and explainable deepfake detec-tion algorithms, valuable for ensuring safe and transparent digital media forensics. | ||
| کلیدواژهها | ||
| Vision Transformer؛ Data-Efficient Image Transformer؛ Face Forgery Detection؛ Ex-plainable AI؛ Transformer Models | ||
| مراجع | ||
|
| ||
|
آمار تعداد مشاهده مقاله: 442 تعداد دریافت فایل اصل مقاله: 533 |
||