Revealing Artifacts via Noise Amplification: A Novel Perspective for AI-Generated Video Detection

2026-06-15

AI Summary

Existing methods struggle to effectively detect AI-generated videos produced by advanced text-to-video models, as these videos exhibit high visual fidelity yet contain subtle distortions in fine details and motion dynamics. This work proposes a novel multi-level noise enhancement framework that, for the first time, integrates bit-plane analysis with a noise amplification mechanism. The approach extracts intrinsic noise from bit planes and synergistically amplifies it at three levels: pixel-wise (intensity enhancement), region-wise (spatial magnification), and frame-wise (temporal aggregation), followed by a lightweight discriminative network for efficient detection. Evaluated on both GenVidBench and a newly introduced HardGVD benchmark, the method significantly outperforms current state-of-the-art techniques, demonstrating strong effectiveness and robustness.

Abstract

With the rapid advancement of video generation models, distinguishing between AI-generated and authentic videos has emerged as a challenging endeavor. The majority of existing research endeavors concentrate on the development of detectors for identifying samples generated by generative adversarial networks. Nevertheless, the detection of AI-generated videos, particularly those produced by text-to-video models, still remains an uncharted territory. Although state-of-the-art text-to-video models can generate realistic visual content similar to real videos, they fall short of generating the details of the images and the changes in details within the videos. Inspired by this, we address AI-generated video detection from a novel perspective of bit-planes, which can effectively describe the details or noises in images or videos. To this end, we propose a simple yet effective approach called Noise Amplification. This approach first extracts noise signals based on bit-planes, then amplifies these noise signals, and finally feeds them into the discriminator networks for video fake classification. Noise amplification is comprehensively constructed by incorporating three aspects: pixel-level intensity enhancement, region-level spatial amplification, and frame-level temporal aggregation. To evaluate methods of AI-generated video detection in challenging scenarios, we also introduce a benchmark named HardGVD. Extensive experiments on both the large-scale dataset GenVidBench and HardGVD show that our simple approach significantly outperforms state-of-the-art methods.

Problem

• AI-generated video detection
• text-to-video models
• video authenticity
• deepfake detection
• synthetic video

Innovation

• Noise Amplification
• Bit-planes
• AI-generated video detection
• Temporal aggregation
• HardGVD