SurFITR: A Dataset for Surveillance Image Forgery Detection and Localisation

TL;DR

SurFITR is a new large-scale dataset designed to improve the detection and localization of subtle, localized forgeries in surveillance images, addressing limitations of existing datasets and models.

Contribution

The paper introduces SurFITR, a comprehensive surveillance image forgery dataset generated with a multimodal pipeline, enabling better training and evaluation of forgery detection methods.

Findings

Existing detectors perform poorly on SurFITR.

Training on SurFITR improves detection accuracy.

SurFITR enhances cross-domain 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.