SpecSem-Net: Integrating Spectral and Semantic Features for Robust AI-generated Video Detection

TL;DR

SpecSem-Net introduces a spectral and semantic feature integration framework with spectral denoising for improved detection of high-fidelity AI-generated videos, addressing limitations of existing methods.

Contribution

It is the first to combine spectral denoising guided by semantic features for robust AI-generated video detection, and provides a new benchmark for evaluation.

Findings

Achieves 87.25% accuracy on the new benchmark.

Achieves 95.59% accuracy on public datasets.

Outperforms existing detection methods.

Abstract

The remarkable visual fidelity of recent commercial video generative models, such as Sora and Veo, renders robust AI-generated video detection increasingly essential to prevent synthetic content from being indistinguishable from real videos and exploited for disinformation. However, existing detectors often fail due to an over-reliance on increasingly realistic semantic features, neglecting subtle spectral artifacts. In this paper, we propose SpecSem-Net, the first framework to introduce a semantic-guided spectral denoising mechanism specifically for high-fidelity AI-generated video detection. Specifically, we design a spectral module to extract high-frequency features via Fourier-Transform based filtering. Furthermore, to reduce misjudgments arising from spectral noise, we employ a Gated Merging Mechanism to adaptively fuse semantic context, effectively mitigating spectral noise. Additionally, to evaluate detector performance on the latest top-tier generative models, we construct a comprehensive benchmark comprising 5 SOTA commercial generators. Extensive experiments demonstrate that SpecSem-Net outperforms existing methods, achieving accuracies of 87.25% and 95.59% on our benchmark and public datasets, respectively.