SurFITR: A Dataset for Surveillance Image Forgery Detection and Localisation

2026-04-08

AI Summary

This work addresses the limited generalization of existing image forgery detection methods in surveillance scenarios, where manipulations are often local, subtle, and semantically complex. To bridge this gap, the authors introduce the first large-scale, fine-grained, and semantics-aware forged image dataset specifically tailored for surveillance contexts. The dataset is constructed by leveraging multimodal large language models to guide diverse image editing tools, generating forensically valuable and semantically consistent manipulation samples. Experimental results demonstrate a significant performance drop of current detectors on this new benchmark, while models trained on the proposed dataset achieve notable improvements both in-domain and in cross-domain settings, effectively addressing the critical data and performance gaps in surveillance-oriented image forgery detection.

Abstract

We present the Surveillance Forgery Image Test Range (SurFITR), a dataset for surveillance-style image forgery detection and localisation, in response to recent advances in open-access image generation models that raise concerns about falsifying visual evidence. Existing forgery models, trained on datasets with full-image synthesis or large manipulated regions in object-centric images, struggle to generalise to surveillance scenarios. This is because tampering in surveillance imagery is typically localised and subtle, occurring in scenes with varied viewpoints, small or occluded subjects, and lower visual quality. To address this gap, SurFITR provides a large collection of forensically valuable imagery generated via a multimodal LLM-powered pipeline, enabling semantically aware, fine-grained editing across diverse surveillance scenes. It contains over 137k tampered images with varying resolutions and edit types, generated using multiple image editing models. Extensive experiments show that existing detectors degrade significantly on SurFITR, while training on SurFITR yields substantial improvements in both in-domain and cross-domain performance. SurFITR is publicly available on GitHub.

Problem

• surveillance image
• image forgery detection
• localisation
• tampering
• generalisation

Innovation

• Surveillance Forgery Detection
• Multimodal LLM
• Fine-grained Image Editing
• Forgery Localization
• Forensic Dataset