Skip to content
Preprint

Not Every Cue Is Evidence: Cue Supportability for Forensic Distillation in Face Forgery Detection

Oct 2026 · 0 citations · 62 references
Computer Science

TL;DR

A dual supportability audit is proposed that filters privileged cues before distillation, preventing unsupported details from leaking into the student and significantly raising explanation faithfulness compared to existing paradigms.

Abstract

Explainable face forgery detection requires not only correct judgments but also forensic evidence supported by the input image. However, training cues are typically derived from the forgery generation pipeline. Since images available to the model are of limited quality and resolution, it is difficult to observe the fine-grained traces these cues refer to. Training methods that further supply such cues as privileged supervision exacerbate the problem, as the student is taught to articulate evidence it cannot perceive. The student consequently resorts to templated or over-generalized expressions, giving rise to the correct-answer, wrong-evidence (CAWE) failure. We propose a dual supportability audit that filters privileged cues before distillation. Each candidate cue must be predictable from the remaining privileged context and causally tied to high-attribution visual regions in the current image. Only cues satisfying both conditions receive substantial distillation weight, preventing unsupported details from leaking into the student. Experiments across eight benchmarks show that our method maintains or improves detection accuracy while significantly raising explanation faithfulness compared to existing paradigms.

View source

Similar papers

Preprint Oct 2026

EviStress: Evidence Stress Learning for Generalizable Face Forgery Detection with Textual Explanations

The rapid spread of face forgeries on the Web challenges the credibility of visual content. In practical Web environments, manipulated media originate from evolving generation pipelines and undergo diverse processing during dissemination, making reliable authenticity assessment under unseen manipulations and visual cha...

Jiu-Yao Jing, Xuanyu Yi, Yu Zheng et al. · 0 citations
#explainable ai Book Open access Nov 2026

GenText-Forensics: Challenge on Explainable Forensics and Adversarial Generation for Text-Centric Images

RealText-V2 is developed and the GenText-Forensics Challenge at ACM Multimedia 2026 is launched to advance reasoning-based text-centric image forensics, and analysis shows that leading systems consistently adopt modular expert pipelines, localization remains the primary performance driver, and evidence conditioning is...

Fan-Wei Zeng, Jian-Shu Li, Chang-Tao Miao et al. · 1 citation
Preprint Oct 2026

Forensic Reserve: Eliciting Latent Knowledge for Image Forgery Detection

As generated images become increasingly realistic, reliable forgery detection is essential for maintaining trust in visual information. However, existing methods primarily rely on task-specific supervision to adapt vision foundation model representations, without fully exploiting internal forensic knowledge to guide de...

Jia-Hua Li, Zixu John, Tom Zhong et al. · 0 citations
Book Open access Sep 2026

Team MSU GenText-Forensics Challenge 2026 Technical Report

A decomposed chain-of-thought (CoT) pipeline that combines a document tampering detector (DTD) with two Qwen3-VL-32B vision-language models, each LoRA-adapted to a distinct sub-task is presented.

Kirill Koltsov, Aleksandr Gushchin, D. Vatolin et al. · 0 citations
Sep 2026

ForensicZoom: Adaptive Visual Inspection with Multimodal LLMs for Industrial-Grade Face Forgery Detection

Reliable face forgery detection is critical to the security of online identity verification systems, where missed attacks compromise security and excessive false positives disrupt legitimate users. Specialized forensic detectors achieve strong detection performance but provide limited interpretability, while multimodal...

Hang Zhou, Yi-Ming Tang, Kun Yu et al. · 0 citations
Book Open access Nov 2026

From Forensic Masks to Ranked Reasons: Evidence-Guided Multimodal Explanation for Document Image Tampering

Document images, including scanned archives, contracts, certificates, and administrative records, are frequently used as digital evidence. Local edits to characters, seals, dates, or handwritten fields can change their meaning while leaving only subtle forensic traces. Existing localization models can predict suspiciou...

Bo Wu, Shu Jiang, Chen-Feng Du et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.