Deepfake Video Detection Using Deep Learning: A Comparative Study with a Novel Attention-CNN-LSTM Architecture

This paper presented a fair multi-model comparison of five deep learning architectures for deepfake video detection on FaceForensics++, including the proposed Attention-CNN-LSTM. The proposed model achieved the highest in-domain accuracy (96.8%) by directing the LSTM toward the most discriminative facial feature channels before temporal reason-ing. An ablation study confirmed the optimal frame sequence length (T =10) and LSTM hidden units (H=256). Statistical tests validated all reported rankings. Cross-dataset evaluation on Celeb-DF revealed that in-domain accuracy and cross-domain generalization do not always align, with EfficientNet-B4 generalizing best despite ranking second in-domain. Key directions for future work include spatial (not just  channel-wise)  attention  for  finer  artifact  localiza- tion; compression-robust training through mixed-quality and simulated-compression augmentation; domain adaptation to improve cross-dataset generalization; and interpretable detec-tion that identifies which facial regions triggered the decision, supporting forensic use cases.

[1]     P. Korshunov and S. Marcel, “Deepfakes: A new threat to face recog-nition? Assessment and detection,” arXiv:1812.08685, 2018. [2]     I. Goodfellow et al., “Generative adversarial networks,” Proc. NeurIPS, vol. 27, 2014. [3]     D. Guera and E. J. Delp, “Deepfake video detection using recurrent neural networks,” Proc. IEEE AVSS, 2018. [4]     D. Afchar, V. Nozick, J. Yamagishi, and I. Echizen, “MesoNet: A compact facial video forgery detection network,” Proc. IEEE WIFS, pp. 1–7, 2018. [5]     A. Rossler et al., “FaceForensics++: Learning to detect manipulated facial images,” Proc. IEEE/CVF ICCV, pp. 1–11, 2019. [6]     F. Chollet, “Xception: Deep learning with depthwise separable convo-lutions,” Proc. IEEE CVPR, pp. 1251–1258, 2017. [7]     Y. Li, M. Chang, and S. Lyu, “In ictu oculi: Exposing AI generated fake face videos by detecting eye blinking,” Proc. IEEE WIFS, 2018. [8]     Y. Li and S. Lyu, “Exposing deepfake videos by detecting face warping artifacts,” Proc. IEEE CVPR Workshops, 2019. [9]     X. Yang, Y. Li, and S. Lyu, “Exposing deep fakes using inconsistent head poses,” Proc. IEEE ICASSP, pp. 8261–8265, 2019. [10]    Y. Li et al., “Celeb-DF: A large-scale challenging dataset for deepfake forensics,” Proc. IEEE/CVF CVPR, pp. 3207–3216, 2020.