DETEKSI DEEPFAKE: BERDASARKAN INKONSISTENSI GERAKAN KEPALA DAN LEHER MENGGUNAKAN VARIATIONAL AUTOENCODER
Keywords:
Deepfake, Variational Autoencoder, Deteksi Anomali, BiomekanikAbstract
Teknologi deepfake saat ini mampu memproduksi manipulasi video yang sangat realistis, sehingga sulit dideteksi oleh metode berbasis spasial statis. Penelitian ini mengusulkan metode deteksi deepfake melalui analisis inkonsistensi biomekanik gerakan kepala dan wajah menggunakan Variational Autoencoder (VAE). Dengan pendekatan unsupervised anomaly detection, model dilatih hanya menggunakan video asli untuk mempelajari pola pergerakan alami manusia. Lima fitur utama diekstraksi menggunakan MediaPipe, yakni Pitch, Yaw, Roll, Eye Aspect Ratio (EAR), dan Mouth Aspect Ratio (MAR), yang diproses dengan teknik sliding window 60 frame. Hasil pengujian menggunakan dataset FaceForensics++ dan Celeb-DF menunjukkan performa model dengan akurasi 66% dan nilai AUC sebesar 0,69. Penelitian ini membuktikan bahwa inkonsistensi gerakan biomekanik merupakan indikator kuat dalam mendeteksi manipulasi video sintesis tingkat lanjut.
References
Jaychand Upadhyay, Vedant Chaudhari, Rupesh Darpe, Rajaram Desai, and Salil Gujar, “Deepfake Detection: A Comprehensive Review of Techniques and Challenges,” Apr. 2025. [Online]. Available: www.ijfmr.com
Yunzhuo Chen, Naveed Akhtar, Nur Al Hasan Haldar, and Ajmal Mian, “Deepfake Detection with Spatio-Temporal Consistency and Attention,” Feb. 2025, [Online]. Available: http://arxiv.org/abs/2502.08216
Sami Belguesmia, Mohand Saïd Allili, and Assia Hamadene, “Unmasking Facial DeepFakes: A Robust Multiview Detection Framework for Natural Images,” Oct. 2025, [Online]. Available: http://arxiv.org/abs/2510.15576
Gan Pei et al., “Deepfake Generation and Detection: A Benchmark and Survey,” ACM Comput. Surv., vol. 58, no. 11, pp. 1–41, Aug. 2026, doi: 10.1145/3801962.
Yuchen Luo, Yong Zhang, Junchi Yan, and Wei Liu, “Generalizing Face Forgery Detection with High-frequency Features,” Mar. 2021, [Online]. Available: http://arxiv.org/abs/2103.12376
Muhammad Ahmad Amin, Yongjian Hu, and Jiankun Hu, “Analyzing temporal coherence for deepfake video detection,” Electronic Research Archive, vol. 32, no. 4, pp. 2621–2641, 2024, doi: 10.3934/ERA.2024119.
Zhihao Gu et al., “Spatiotemporal Inconsistency Learning for DeepFake Video Detection,” in MM 2021 - Proceedings of the 29th ACM International Conference on Multimedia, Association for Computing Machinery, Inc, Oct. 2021, pp. 3473–3481. doi: 10.1145/3474085.3475508.
O’Shaughnessy, Andrea Prati, Cosmina-Mihaela Rosca, and Adrian Stancu, “AI Anomaly-Based Deepfake Detection Using Customized Mahalanobis Distance and Head Pose with Facial Landmarks,” 2025, doi: 10.3390/app.
Byeong Seon An, Hyeji Lim, Hyeon Ah Seong, and Eui Chul Lee, “Facial and Neck Region Analysis for Deepfake Detection Using Remote Photoplethysmography Signal Similarity,” IET Biom., vol. 2024, no. 1, 2024, doi: 10.1049/bme2/7095412.
Saud Sohail, Syed Muhammad Sajjad, Adeerl Zafar, Zafar Iqbal, Zia Muhammad, and Muhammad Kazim, “Deepfake Detection Using Deep Learning: A Unified Forensic Approach to Detect AI-Generated Images and Videos with Fusion of Eye, Nose, and Mouth Landmarks,” Feb. 07, 2025. doi: 10.20944/preprints202502.0552.v1.
Huy Hoang Nguyen, Cuong Nhat Nguyen, Xuan Tung Dao, Quoc Trung Duong, Dzung Pham Thi Kim, and Minh-Tan Pham, “Variational Autoencoder for Anomaly Detection: A Comparative Study,” Aug. 2024, [Online]. Available: http://arxiv.org/abs/2408.13561
Rongju Yao, Zhiqing Bai, Jing Tong, and Khosro Rezaee, “Deepfake Detection in Image Sequences: A Temporal Approach for Anomaly Detection,” International Journal of Intelligent Systems, vol. 2025, no. 1, 2025, doi: 10.1155/int/8566328.
Yu Tian et al., “Unsupervised Anomaly Detection in Medical Images with a Memory-augmented Multi-level Cross-attentional Masked Autoencoder,” Aug. 2023, [Online]. Available: http://arxiv.org/abs/2203.11725
Kurniawan Nur Ramadhani, Rinaldi Munir, and Nugraha Priya Utama, “Improving Video Vision Transformer for Deepfake Video Detection Using Facial Landmark, Depthwise Separable Convolution and Self Attention,” IEEE Access, vol. 12, pp. 8932–8939, 2024, doi: 10.1109/ACCESS.2024.3352890.
Yuting Xu, Jian Liang, Lijun Sheng, and Xiao-Yu Zhang, “Learning Spatiotemporal Inconsistency via Thumbnail Layout for Face Deepfake Detection,” Mar. 2024, [Online]. Available: http://arxiv.org/abs/2403.10261
Chao Feng, Ziyang Chen, and Andrew Owens, “Self-Supervised Video Forensics by Audio-Visual Anomaly Detection,” Mar. 2023, [Online]. Available: http://arxiv.org/abs/2301.01767
Yang Wang, Tianying Liu, Jiaogen Zhou, and Jihong Guan, “Video anomaly detection based on spatio-temporal relationships among objects,” Neurocomputing, vol. 532, pp. 141–151, May 2023, doi: 10.1016/j.neucom.2023.02.027.
Elisabeth Tchaptchet, Elie Fute Tagne, Jaime Acosta, Danda B. Rawat, and Charles Kamhoua, “Deepfakes Detection by Iris Analysis,” IEEE Access, vol. 13, pp. 8977–8987, Jan. 2025, doi: 10.1109/ACCESS.2025.3527868.
Lord Sen and Shyamapada Mukherjee, “A Novel Unified Approach to Deepfake Detection,” Jan. 2026, [Online]. Available: http://arxiv.org/abs/2601.03382
Kamendra Prakash Rahatwal, Sumit Pundir, Mohammad Wazid, and Vivekananda Bhat K., “A Novel Approach to Deepfake Detection: Leveraging Fused Facial and Body Dynamics With a CNN–Transformer Hybrid Network,” IEEE Access, vol. 13, pp. 197085–197108, 2025, doi: 10.1109/ACCESS.2025.3632155.
Deressa Wodajo Deressa, Hannes Mareen, Peter Lambert, Solomon Atnafu, Zahid Akhtar, and Glenn Van Wallendael, “GenConViT: Deepfake Video Detection Using Generative Convolutional Vision Transformer,” Applied Sciences (Switzerland), vol. 15, no. 12, Jun. 2025, doi: 10.3390/app15126622.
Siddiqui Muhammad Yasir and Hyun Kim, “Lightweight Deepfake Detection Based on Multi-Feature Fusion,” Feb. 2025, doi: 10.3390/app15041954.
