Search papers, labs, and topics across Lattice.
This paper introduces Hamiltonian Action Anomaly Detection (HAAD), a novel deepfake detection method based on the hypothesis that real images reside in stable, low-energy states on the image latent manifold, while deepfakes occupy unstable, high-energy states. HAAD uses Hamiltonian-inspired dynamics to probe image stability, quantifying dynamic behaviors through trajectory statistics like Hamiltonian action and energy dissipation. Experiments demonstrate that HAAD outperforms state-of-the-art baselines in cross-dataset transfer scenarios, suggesting a more robust approach to detecting evolving deepfake techniques.
Escaping the endless cat-and-mouse game of deepfake detection may be possible by shifting from static pattern recognition to physics-inspired dynamical stability analysis, where real images are stable and deepfakes are not.
Driven by the rapid development of generative AI models, deepfake detectors are compelled to undergo periodic recalibration to capture newly developed synthetic artifacts. To break this cycle, we propose a new perspective on deepfake detection: moving from static pattern recognition to dynamical stability analysis. Specifically, our approach is motivated by physics-inspired priors: we hypothesize that natural images, as products of dissipative physical processes, tend to settle near stable, low-energy equilibria. In contrast, generative models optimize for statistical similarity to real images but do not explicitly enforce structural constraints such as geometric smoothness, leaving deepfakes more likely to occupy unstable, high-energy states. To operationalize this, we introduce Hamiltonian Action Anomaly Detection (HAAD), comprising three contributions: \textbf{i)} We model the image latent manifold as a potential energy surface. Under this hypothesis, real images are expected to produce basin-like low-energy responses, whereas fake images are more likely to induce high-potential, high-gradient responses. \textbf{ii)} We employ Hamiltonian-inspired dynamics as a stability probe. By releasing latent states from rest, samples near stable regions remain bounded, while high-gradient samples produce larger trajectory responses. \textbf{iii)} We quantify these dynamic behaviors through two trajectory statistics, \ie, Hamiltonian action and energy dissipation. Extensive experiments show that HAAD outperforms evaluated state-of-the-art baselines on challenging cross-dataset transfer benchmarks, supporting a physics-inspired stability prior for digital forensics.
