Deepfake Video Forgery Detection

References

[1] Chih-Chung Hsu, Tzu-Yi Hung, Chia-Wen Lin and Chiou-Ting Hsu, "Video forgery detection using correlation of noise residue," 2008 IEEE 10th Workshop on Multimedia Signal Processing, Cairns, Qld, 2008, pp. 170-174, doi: 10.1109/MMSP.2008.4665069.

[2] D. Afchar, V. Nozick, J. Yamagishi and I. Echizen, "MesoNet: a Compact Facial Video Forgery Detection Network," 2018 IEEE International Workshop on Information Forensics and Security (WIFS), Hong Kong, China, 2018, pp. 1-7, doi: 10.1109/WIFS.2018.8630761.

[3] Shelke, N.A., Kasana, S.S. A comprehensive survey on passive techniques for digital video forgery detection. Multimed Tools Appl 80, 6247–6310 (2021). https://doi.org/10.1007/s11042-020-09974-4

[4] A. W. A. Wahab, M. A. Bagiwa, M. Y. I. Idris, S. Khan, Z. Razak and M. R. K. Ariffin, "Passive video forgery detection techniques: A survey," 2014 10th International Conference on Information Assurance and Security, Okinawa, Japan, 2014, pp. 29-34, doi: 10.1109/ISIAS.2014.7064616.

[5] A. V. Subramanyam and S. Emmanuel, "Video forgery detection using HOG features and compression properties," 2012 IEEE 14th International Workshop on Multimedia Signal Processing (MMSP), Banff, AB, Canada, 2012, pp. 89-94, doi: 10.1109/MMSP.2012.6343421.

[6] M. Aloraini, M. Sharifzadeh and D. Schonfeld, "Sequential and Patch Analyses for Object Removal Video Forgery Detection and Localization," in IEEE Transactions on Circuits and Systems for Video Technology, vol. 31, no. 3, pp. 917-930, March 2021, doi: 10.1109/TCSVT.2020.2993004

[7] Y. Li, X. Yang, P. Sun, H. Qi and S. Lyu, "Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics," 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA, 2020, pp. 3204-3213, doi: 10.1109/CVPR42600.2020.00327.

[8] Deng L, Suo H, Li D. Deepfake Video Detection Based on EfficientNet-V2 Network. Comput Intell Neurosci. 2022 Apr 15;2022:3441549. doi: 10.1155/2022/3441549. PMID: 35463269; PMCID: PMC9033321.

[9] Stehouwer J, Dang H, Liu F, Liu X, Jain A. On the detection of digital face manipulation. arXiv preprint. arXiv:1910.01717 (2019)

[10] Rahmouni N, Nozick V, Yamagishi J, Echizen I. Distinguishing computer graphics from natural images using convolution neural networks. In: 2017 IEEE workshop on information forensics and security (WIFS). IEEE; 2017. p. 1–6.

[11] Jing Zhang, Yuting Su, and Mingyu Zhang. 2009. Exposing digital video forgery by ghost shadow artifact. In Proceedings of the First ACM workshop on Multimedia in forensics (MiFor '09). Association for Computing Machinery, New York, NY, USA, 49–54. https://doi.org/10.1145/1631081.1631093

@article{1704725,
      author = {Apurv Jindal},
      title = {Deepfake Video Forgery Detection},
      journal = {Iconic Research And Engineering Journals},
      year = {2023},
      volume = {6},
      number = {12},
      pages = {765-768},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1704725.pdf},
      abstract = {With the proliferation of digital video content and the advancements in video editing tools, video forgery has become a critical concern in various domains, including journalism, surveillance, and legal proceedings. Detecting forged videos is a challenging task due to the increasing sophistication of forgery techniques. Traditional techniques include forensic watermarking, temporal inconsistencies analysis, and sensor pattern noise analysis. This research paper proposes a novel deepfake detection approach combining ResNet-50 and LSTM networks, a deep learning architecture known for its excellent performance in image recognition tasks. The proposed method leverages the strengths of both spatial and temporal modeling to enhance the detection accuracy. The ResNet-50 model is utilized to extract spatial features from individual frames, capturing visual cues and inconsistencies introduced by deepfake manipulations. The LSTM network is employed to model the temporal dependencies between frames, enabling the detection of subtle temporal artifacts that may indicate the presence of deepfakes.
To train the model, a large-scale dataset comprising both real and deepfake videos is utilized. The dataset is carefully curated, ensuring a diverse range of deepfake manipulations and real-world scenarios. The ResNet-50 backbone is pre-trained on a large image dataset, allowing it to learn generic visual representations that are then fine-tuned for deepfake detection. The LSTM network is trained to capture the temporal dynamics and patterns specific to deepfake videos.},
      keywords = {CNN, Deepfake Detection, LSTM, ResNet50, Video Forgery Detection},
      month = {June},
  }