Research

Paper

TESTING February 25, 2026

Beyond Static Artifacts: A Forensic Benchmark for Video Deepfake Reasoning in Vision Language Models

Authors

Zheyuan Gu, Qingsong Zhao, Yusong Wang, Zhaohong Huang, Xinqi Li, Cheng Yuan, Jiaowei Shao, Chi Zhang, Xuelong Li

Abstract

Current Vision-Language Models (VLMs) for deepfake detection excel at identifying spatial artifacts but overlook a critical dimension: temporal inconsistencies in video forgeries. Adapting VLMs to reason about these dynamic cues remains a distinct challenge. To bridge this gap, we propose Forensic Answer-Questioning (FAQ), a large-scale benchmark that formulates temporal deepfake analysis as a multiple-choice task. FAQ introduces a three-level hierarchy to progressively evaluate and equip VLMs with forensic capabilities: (1) Facial Perception, testing the ability to identify static visual artifacts; (2) Temporal Deepfake Grounding, requiring the localization of dynamic forgery artifacts across frames; and (3) Forensic Reasoning, challenging models to synthesize evidence for final authenticity verdicts. We evaluate a range of VLMs on FAQ and generate a corresponding instruction-tuning set, FAQ-IT. Extensive experiments show that models fine-tuned on FAQ-IT achieve advanced performance on both in-domain and cross-dataset detection benchmarks. Ablation studies further validate the impact of our key design choices, confirming that FAQ is the driving force behind the temporal reasoning capabilities of these VLMs.

Metadata

arXiv ID: 2602.21779
Provider: ARXIV
Primary Category: cs.CV
Published: 2026-02-25
Fetched: 2026-02-26 05:00

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.21779v1</id>\n    <title>Beyond Static Artifacts: A Forensic Benchmark for Video Deepfake Reasoning in Vision Language Models</title>\n    <updated>2026-02-25T10:54:55Z</updated>\n    <link href='https://arxiv.org/abs/2602.21779v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.21779v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Current Vision-Language Models (VLMs) for deepfake detection excel at identifying spatial artifacts but overlook a critical dimension: temporal inconsistencies in video forgeries. Adapting VLMs to reason about these dynamic cues remains a distinct challenge. To bridge this gap, we propose Forensic Answer-Questioning (FAQ), a large-scale benchmark that formulates temporal deepfake analysis as a multiple-choice task. FAQ introduces a three-level hierarchy to progressively evaluate and equip VLMs with forensic capabilities: (1) Facial Perception, testing the ability to identify static visual artifacts; (2) Temporal Deepfake Grounding, requiring the localization of dynamic forgery artifacts across frames; and (3) Forensic Reasoning, challenging models to synthesize evidence for final authenticity verdicts. We evaluate a range of VLMs on FAQ and generate a corresponding instruction-tuning set, FAQ-IT. Extensive experiments show that models fine-tuned on FAQ-IT achieve advanced performance on both in-domain and cross-dataset detection benchmarks. Ablation studies further validate the impact of our key design choices, confirming that FAQ is the driving force behind the temporal reasoning capabilities of these VLMs.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CV'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.AI'/>\n    <published>2026-02-25T10:54:55Z</published>\n    <arxiv:comment>16 pages, 9 figures. Submitted to CVPR 2026</arxiv:comment>\n    <arxiv:primary_category term='cs.CV'/>\n    <author>\n      <name>Zheyuan Gu</name>\n    </author>\n    <author>\n      <name>Qingsong Zhao</name>\n    </author>\n    <author>\n      <name>Yusong Wang</name>\n    </author>\n    <author>\n      <name>Zhaohong Huang</name>\n    </author>\n    <author>\n      <name>Xinqi Li</name>\n    </author>\n    <author>\n      <name>Cheng Yuan</name>\n    </author>\n    <author>\n      <name>Jiaowei Shao</name>\n    </author>\n    <author>\n      <name>Chi Zhang</name>\n    </author>\n    <author>\n      <name>Xuelong Li</name>\n    </author>\n  </entry>"
}