Interpretable facial dynamics as behavioral and perceptual traces of deepfakes
Authors: Timothy Joseph Murphy, Jennifer Cook, Hélio Clemente José Cuve
Organizations: School of Psychology, University of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom · School of Psychology and Neuroscience, University of Bristol, 12a Priory Road, Bristol, BS8 1TU, United Kingdom
Abstract
Deepfake detection research has largely converged on deep learning approaches that, despite strong benchmark performance, offer limited insight into what distinguishes real from manipulated facial behavior. This study presents an interpretable alternative grounded in bio-behavioral features of facial dynamics and evaluates how computational detection strategies relate to human perceptual judgments. We identify core low-dimensional patterns of facial movement, from which temporal features characterizing spatiotemporal structure were derived. Traditional machine learning classifiers trained on these features achieved modest but significant above-chance deepfake classification, driven by higher-order temporal irregularities that were more pronounced in manipulated than real facial dynamics. Notably, detection was substantially more accurate for videos containing emotive expressions than those without. An emotional valence classification analysis further indicated that emotive signals are systematically degraded in deepfakes, explaining the differential impact of emotive dynamics on detection. Furthermore, we provide an additional and often overlooked dimension of explainability by assessing the relationship between model decisions and human perceptual detection. Model and human judgments converged for emotive but diverged for non-emotive videos, and even where outputs aligned, underlying detection strategies differed. These findings demonstrate that face-swapped deepfakes carry a measurable behavioral fingerprint, most salient during emotional expression. Additionally, model-human comparisons suggest that interpretable computational features and human perception may offer complementary rather than redundant routes to detection.
With every advancement in generative AI models, forensics is under increasing pressure. The constant emergence of new generation techniques makes it impossible to collect data for each manipulation to train a deepfake detection model. Thus, generalizing to deepfakes unseen during training is one of the major challenges in current deepfake detection research. To tackle this challenge, we employ high-level semantic cues and argue that these cues can support low-level focused approaches in generalizing to unseen types of manipulations. In this work, we study emotions as a high-level semantic cue. We propose Emo-Boost, a multimodal deepfake detection framework that fuses an off-the-shelf RGB- and acoustic-focused deepfake detector with our emotion-based deepfake detector EmoForensics. EmoForensics utilises vision and audio emotion recognition modules and models intra- and inter-modal temporal consistency in emotion representations from an audio-visual stream. We found that EmoForensics and the low-level focused method capture complementary signals. Consequently, combining both signals in EmoBoost enhances the average cross-manipulation generalization AUC by 2.1% on FakeAVCeleb.
Deepfake detectors show large performance gaps across demographic groups. Existing fairness approaches require demographic labels, retraining, or sacrifice accuracy. We introduce Face-Fairness (FF), a plug-and-play framework for bias mitigation. Our primary contribution, Face-Feature Tuning (FFT), is the first demographic label-free fairness method demonstrated for deepfake detection: a lightweight calibrator that performs a logit remapping conditioned on frozen face embeddings. We complement FFT with two variants: FF-Max, which maximizes worst-group accuracy when demographics are available, and FF-Discover, which does the same with embedding-discovered groups. Across in-domain and cross-dataset test settings, FF consistently reduces FPR/TPR gaps and improves minimum group accuracy while maintaining (often improving) overall accuracy. The approach is detector-agnostic, adds negligible runtime overhead, and requires no access to identity attributes.
Recent deepfake detection methods demonstrate improved cross-dataset generalization, yet the underlying mechanisms remain underexplored. We introduce the Alpha Blending Hypothesis, positing that state-of-the-art frame-based detectors primarily function as alpha blending searchers; rather than learning semantic anomalies or specific generative neural fingerprints, they localize low-level compositing artifacts introduced during the integration of manipulated faces into target frames. We experimentally validate the hypothesis, demonstrating that deepfake detectors exhibit high sensitivity to the so-called self-blended images (SBI) and non-generative manipulations. We propose the method BlenD that leverages a large-scale, diverse dataset of real-only facial images augmented with SBI. This approach achieves the best average cross-dataset generalization on 15 compositional deepfake datasets released between 2019 and 2025 without utilizing explicitly generated deepfakes during training. Furthermore, we show that predictions from explicit blending searchers and models resilient to blending shortcuts are highly complementary, yielding a state-of-the-art AUROC of 94.0% in an ensemble configuration. The code with experiments and the trained model will be publicly released.