Micro-Expression-Aware Avatar Fingerprinting via Inter-Frame Feature Differencing

2026-04-25

AI Summary

This work addresses the challenge of verifying speaker identity—rather than merely detecting authenticity—in synthesized talking-head videos. It proposes an end-to-end, preprocessing-free method for extracting identity-specific motion fingerprints directly from raw video sequences. Departing from conventional non-differentiable landmark-based pipelines, the approach innovatively integrates an F5C micro-expression-aware backbone with an inter-frame deep feature differencing mechanism to capture motion-dominated identity-related dynamics while effectively suppressing static appearance cues. Evaluated on the NVFAIR dataset, the method achieves an AUC of 0.877 and consistently matches or outperforms landmark-based baselines across most cross-generator settings.

Technology Category

• Computer Vision: Motion & Tracking
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Avatar fingerprinting, i.e., verifying who drives a synthetic talking-head video rather than whether it is real, is a critical safeguard for authorized use of face-reenactment technology. Existing methods rely on a fixed, non-differentiable landmark extraction stage that prevents the fingerprinting model from being optimized end-to-end from raw pixels. We propose a preprocessing-free system built on a micro-expression-aware backbone operating on raw video frames, with inter-frame feature differencing as the core design principle: consecutive feature maps are subtracted in the learned deep feature space, so that temporally stable appearance dimensions contribute zero to the output while driver-specific motion dynamics are preserved. A controlled ablation on NVFAIR confirms that temporal motion accounts for the large majority of discriminative performance, and that raw appearance features actively degrade identity separation. Both the choice of backbone and the differencing principle are essential: differencing alone is insufficient when applied to a generic encoder, as appearance-dominated features collapse to near-identical representations across adjacent frames, while the micro-expression-aware F5C backbone retains measurable motion variation that the differencing operation can exploit. Without any external preprocessing, our model achieves an overall AUC of 0.877 on NVFAIR and matches or exceeds the landmark-based baseline on the majority of cross-generator pairs.

Problem

• avatar fingerprinting
• micro-expression
• inter-frame feature differencing
• face reenactment
• identity verification

Innovation

• micro-expression-aware
• inter-frame feature differencing
• avatar fingerprinting
• end-to-end optimization
• motion dynamics