A Lightweight and Effective Image Tampering Localization Network with Vision Mamba

2025-02-14

AI Summary

Existing image forgery localization methods face a fundamental trade-off between global modeling capability (limited by CNNs) and computational efficiency (hampered by Transformer’s quadratic complexity). Method: We propose ForMa—the first lightweight localization network for blind image forensics that integrates Vision Mamba, a linear-complexity state-space model, to achieve efficient global dependency modeling. It introduces a parameter-free pixel-rearrangement upsampling decoder and a novel noise-augmented decoding mechanism to enhance sensitivity to subtle tampering cues. Multi-scale Vision Mamba features are fused to capture hierarchical long-range contextual information. Contributions/Results: (1) First application of Vision Mamba to image forgery localization; (2) Novel noise-injection decoding and parameter-free upsampling strategy. ForMa achieves state-of-the-art generalization and robustness across 10 standard benchmarks with the lowest computational overhead. Code is publicly available.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Cyber-crime defenses and forensics
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Current image tampering localization methods primarily rely on Convolutional Neural Networks (CNNs) and Transformers. While CNNs suffer from limited local receptive fields, Transformers offer global context modeling at the expense of quadratic computational complexity. Recently, the state space model Mamba has emerged as a competitive alternative, enabling linear-complexity global dependency modeling. Inspired by it, we propose a lightweight and effective FORensic network based on vision MAmba (ForMa) for blind image tampering localization. Firstly, ForMa captures multi-scale global features that achieves efficient global dependency modeling through linear complexity. Then the pixel-wise localization map is generated by a lightweight decoder, which employs a parameter-free pixel shuffle layer for upsampling. Additionally, a noise-assisted decoding strategy is proposed to integrate complementary manipulation traces from tampered images, boosting decoder sensitivity to forgery cues. Experimental results on 10 standard datasets demonstrate that ForMa achieves state-of-the-art generalization ability and robustness, while maintaining the lowest computational complexity. Code is available at https://github.com/multimediaFor/ForMa.

Problem

• Lightweight image tampering localization
• Global dependency modeling with linear complexity
• Enhanced decoder sensitivity to forgery cues

Innovation

• Vision Mamba for global modeling
• Linear-complexity feature capture
• Noise-assisted decoding strategy