# Beyond Ambiguous Visual Cues: Studying Physiological Disruptions and Cross-Modal Inconsistencies in Deepfake Videos

arXiv: https://arxiv.org/abs/2609.12668 · HTML: https://mlsanity.com/paper/2609.12668/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 37 references that are in the map.
- First posted on arXiv 11 Sep 2026.

## Builds on (most cited first, 24 of 24)

- [Analyzing and Improving the Image Quality of StyleGAN](https://mlsanity.com/paper/1912.04958/) (2019)
- [FaceForensics++: Learning to Detect Manipulated Facial Images](https://mlsanity.com/paper/1901.08971/) (2019)
- [Celeb-DF: A Large-scale Challenging Dataset for DeepFake Forensics](https://mlsanity.com/paper/1909.12962/) (2019)
- [Face X-ray for More General Face Forgery Detection](https://mlsanity.com/paper/1912.13458/) (2019)
- [One-Shot Free-View Neural Talking-Head Synthesis for Video Conferencing](https://mlsanity.com/paper/2011.15126/) (2020)
- [MesoNet: a Compact Facial Video Forgery Detection Network](https://mlsanity.com/paper/1809.00888/) (2018)
- [FSGAN: Subject Agnostic Face Swapping and Reenactment](https://mlsanity.com/paper/1908.05932/) (2019)
- [Exposing DeepFake Videos By Detecting Face Warping Artifacts](https://mlsanity.com/paper/1811.00656/) (2018)
- [Exposing Deep Fakes Using Inconsistent Head Poses](https://mlsanity.com/paper/1811.00661/) (2018)
- [Lips Don't Lie: A Generalisable and Robust Approach to Face Forgery Detection](https://mlsanity.com/paper/2012.07657/) (2020)
- [SimSwap: An Efficient Framework For High Fidelity Face Swapping](https://mlsanity.com/paper/2106.06340/) (2021)
- [FaceShifter: Towards High Fidelity And Occlusion Aware Face Swapping](https://mlsanity.com/paper/1912.13457/) (2019)
- [Thin-Plate Spline Motion Model for Image Animation](https://mlsanity.com/paper/2203.14367/) (2022)
- [LivePortrait: Efficient Portrait Animation with Stitching and Retargeting Control](https://mlsanity.com/paper/2407.03168/) (2024)
- [FakeCatcher: Detection of Synthetic Portrait Videos using Biological Signals](https://mlsanity.com/paper/1901.02212/) (2019)
- [DiffFace: Diffusion-based Face Swapping with Facial Guidance](https://mlsanity.com/paper/2212.13344/) (2022)
- [On Face Segmentation, Face Swapping, and Face Perception](https://mlsanity.com/paper/1704.06729/) (2017)
- [Deepfake Generation and Detection: A Benchmark and Survey](https://mlsanity.com/paper/2403.17881/) (2024)
- [DeepRhythm: Exposing DeepFakes with Attentional Visual Heartbeat Rhythms](https://mlsanity.com/paper/2006.07634/) (2020)
- [rPPG-Toolbox: Deep Remote PPG Toolbox](https://mlsanity.com/paper/2210.00716/) (2022)
- [In Ictu Oculi: Exposing AI Generated Fake Face Videos by Detecting Eye Blinking](https://mlsanity.com/paper/1806.02877/) (2018)
- [A Reproducible Study on Remote Heart Rate Measurement](https://mlsanity.com/paper/1709.00962/) (2017)
- [DeepFakesON-Phys: DeepFakes Detection based on Heart Rate Estimation](https://mlsanity.com/paper/2010.00400/) (2020)
- [Motion Matters: Neural Motion Transfer for Better Camera Physiological Measurement](https://mlsanity.com/paper/2303.12059/) (2023)

## Abstract (arXiv)

Recent deepfake detection studies increasingly suggest remote photoplethysmography (rPPG) signals as an authenticity cue. However, existing benchmarks lack physiological ground truth, and current detectors underexplore the cross-level relationship between facial features and physiological dynamics, often relying on late fusion or rPPG features alone. In this paper, we construct high-fidelity deepfake manipulations on established real rPPG datasets (COHFACE and UBFC-rPPG) to investigate how forgeries disrupt natural physiological signals and facial behavior at the same time. Building on this analysis, we propose a bidirectional co-attention fusion detector that jointly models rPPG and facial behavior tokens. This mechanism explicitly captures the cross-level dependencies between pulse dynamics and facial motion to learn a robust, joint authenticity representation. Extensive experiments using a subject-disjoint 5-fold evaluation demonstrate the superiority of our approach. Achieving a 92.80% AUC on constructed datasets using face swapping and 96.78% AUC on motion transfer, our model outperforms both the rPPG-only single modality baseline and the best feature-level fusion methods. Furthermore, transfer-learning result of the fusion detector on Celeb-DF-v2 while keeping both feature extractors fixed achieves 91.20% accuracy and 86.08% AUC, which suggests applicability under target-domain adaptation.