# T2IShield: Defending Against Backdoors on Text-to-Image Diffusion Models

arXiv: https://arxiv.org/abs/2407.04215 · HTML: https://mlsanity.com/paper/2407.04215/

## Key facts

- Cited by 29 papers in the map (data through 14 Oct 2026).
- Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models.
- Builds on 19 of its 54 references that are in the map.
- First posted on arXiv 5 Jul 2024.

## Placed in

- Diffusion models / [Robustness, privacy and adversarial aspects](https://mlsanity.com/diffusion/robustness-privacy-and-adversarial-aspects/)
- LLM safety and security / [Backdoor and data poisoning attacks](https://mlsanity.com/safety/backdoor-and-data-poisoning-attacks/)

## Builds on (most cited first, 19 of 19)

- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Prompt-to-Prompt Image Editing with Cross Attention Control](https://mlsanity.com/paper/2208.01626/) (2022)
- [Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networks](https://mlsanity.com/paper/1805.12185/) (2018)
- [Detecting Backdoor Attacks on Deep Neural Networks by Activation Clustering](https://mlsanity.com/paper/1811.03728/) (2018)
- [Unified Concept Editing in Diffusion Models](https://mlsanity.com/paper/2308.14761/) (2023)
- [DiffusionDB: A Large-scale Prompt Gallery Dataset for Text-to-Image Generative Models](https://mlsanity.com/paper/2210.14896/) (2022)
- [Reflection Backdoor: A Natural Backdoor Attack on Deep Neural Networks](https://mlsanity.com/paper/2007.02343/) (2020)
- [Forget-Me-Not: Learning to Forget in Text-to-Image Diffusion Models](https://mlsanity.com/paper/2303.17591/) (2023)
- [StableVITON: Learning Semantic Correspondence with Latent Diffusion Model for Virtual Try-On](https://mlsanity.com/paper/2312.01725/) (2023)
- [TABOR: A Highly Accurate Approach to Inspecting and Restoring Trojan Backdoors in AI Systems](https://mlsanity.com/paper/1908.01763/) (2019)
- [VillanDiffusion: A Unified Backdoor Attack Framework for Diffusion Models](https://mlsanity.com/paper/2306.06874/) (2023)
- [Towards Safe Self-Distillation of Internet-Scale Text-to-Image Diffusion Models](https://mlsanity.com/paper/2307.05977/) (2023)
- [BAGM: A Backdoor Attack for Manipulating Text-to-Image Generative Models](https://mlsanity.com/paper/2307.16489/) (2023)
- [ReFACT: Updating Text-to-Image Models by Editing the Text Encoder](https://mlsanity.com/paper/2306.00738/) (2023)
- [DisDet: Exploring Detectability of Backdoor Attack on Diffusion Models](https://mlsanity.com/paper/2402.02739/) (2024)
- [Backdooring Textual Inversion for Concept Censorship](https://mlsanity.com/paper/2308.10718/) (2023)
- [Can There be Art Without an Artist?](https://mlsanity.com/paper/2209.07667/) (2022)

## Built on by (most cited first, 29 of 29)

- [SAFREE: Training-Free and Adaptive Guard for Safe Text-to-Image And Video Generation](https://mlsanity.com/paper/2410.12761/) (2024)
- [Defending Text-to-image Diffusion Models: Surprising Efficacy of Textual Perturbations Against Backdoor Attacks](https://mlsanity.com/paper/2408.15721/) (2024)
- [Erased but Not Forgotten: How Backdoors Compromise Concept Erasure](https://mlsanity.com/paper/2504.21072/) (2025)
- [BackdoorDM: A Comprehensive Benchmark for Backdoor Learning on Diffusion Model](https://mlsanity.com/paper/2502.11798/) (2025)
- [Efficient Input-level Backdoor Defense on Text-to-Image Synthesis via Neuron Activation Variation](https://mlsanity.com/paper/2503.06453/) (2025)
- [Trustworthy Text-to-Image Diffusion Models: A Timely and Focused Survey](https://mlsanity.com/paper/2409.18214/) (2024)
- [CopyrightShield: Enhancing Diffusion Model Security against Copyright Infringement Attacks](https://mlsanity.com/paper/2412.01528/) (2024)
- [MixBridge: Heterogeneous Image-to-Image Backdoor Attack through Mixture of Schr"odinger Bridges](https://mlsanity.com/paper/2505.08809/) (2025)
- [When LoRA Betrays: Backdooring Text-to-Image Models by Masquerading as Benign Adapters](https://mlsanity.com/paper/2602.21977/) (2026)
- [A Dual-Purpose Framework for Backdoor Defense and Backdoor Amplification in Diffusion Models](https://mlsanity.com/paper/2502.19047/) (2025)
- [Trigger without Trace: Towards Stealthy Backdoor Attack on Text-to-Image Diffusion Models](https://mlsanity.com/paper/2503.17724/) (2025)
- [Responsible Diffusion: A Comprehensive Survey on Safety, Ethics, and Trust in Diffusion Models](https://mlsanity.com/paper/2509.22723/) (2025)
- [Combinational Backdoor Attack against Customized Text-to-Image Models](https://mlsanity.com/paper/2411.12389/) (2024)
- [On the Fairness, Diversity and Reliability of Text-to-Image Generative Models](https://mlsanity.com/paper/2411.13981/) (2024)
- [Latent Beam Diffusion Models for Generating Visual Sequences](https://mlsanity.com/paper/2503.20429/) (2025)
- [Parasite: A Steganography-based Backdoor Attack Framework for Diffusion Models](https://mlsanity.com/paper/2504.05815/) (2025)
- [Towards Dataset Copyright Evasion Attack against Personalized Text-to-Image Diffusion Models](https://mlsanity.com/paper/2505.02824/) (2025)
- [ME: Trigger Element Combination Backdoor Attack on Copyright Infringement](https://mlsanity.com/paper/2506.10776/) (2025)
- [Gungnir: Exploiting Stylistic Features in Images for Backdoor Attacks on Diffusion Models](https://mlsanity.com/paper/2502.20650/) (2025)
- [Practical, Generalizable and Robust Backdoor Attacks on Text-to-Image Diffusion Models](https://mlsanity.com/paper/2508.01605/) (2025)
- [BadRSSD: Backdoor Attacks on Regularized Self-Supervised Diffusion Models](https://mlsanity.com/paper/2603.01019/) (2026)
- [BlackMirror: Black-Box Backdoor Detection for Text-to-Image Models via Instruction-Response Deviation](https://mlsanity.com/paper/2603.05921/) (2026)
- [Token by Token, Compromised: Backdoor Vulnerabilities in Unified Autoregressive Models](https://mlsanity.com/paper/2605.19227/) (2026)
- [BadBlocks: Low-Cost and Stealthy Backdoor Attacks Tailored for Text-to-Image Diffusion Models](https://mlsanity.com/paper/2508.03221/) (2025)
- [Backdoor Sentinel: Detecting and Detoxifying Backdoors in Diffusion Models via Temporal Noise Consistency](https://mlsanity.com/paper/2602.01765/) (2026)
- [Semantic-level Backdoor Attack against Text-to-Image Diffusion Models](https://mlsanity.com/paper/2602.04898/) (2026)
- [Beyond the False Trade-off: Adaptive EWC for Stealthy and Generalizable T2I Backdoors](https://mlsanity.com/paper/2605.08280/) (2026)
- [Awakening the Hydra: Stabilizing Multi-Concept Backdoor Injection in Text-to-Image Diffusion Models](https://mlsanity.com/paper/2605.19698/) (2026)
- [BadAction: Backdoor Attacks on Interactive Video Generation via Action-Guided Triggers](https://mlsanity.com/paper/2609.39047/) (2026)

## Abstract (arXiv)

While text-to-image diffusion models demonstrate impressive generation capabilities, they also exhibit vulnerability to backdoor attacks, which involve the manipulation of model outputs through malicious triggers. In this paper, for the first time, we propose a comprehensive defense method named T2IShield to detect, localize, and mitigate such attacks. Specifically, we find the "Assimilation Phenomenon" on the cross-attention maps caused by the backdoor trigger. Based on this key insight, we propose two effective backdoor detection methods: Frobenius Norm Threshold Truncation and Covariance Discriminant Analysis. Besides, we introduce a binary-search approach to localize the trigger within a backdoor sample and assess the efficacy of existing concept editing methods in mitigating backdoor attacks. Empirical evaluations on two advanced backdoor attack scenarios show the effectiveness of our proposed defense method. For backdoor sample detection, T2IShield achieves a detection F1 score of 88.9% with low computational cost. Furthermore, T2IShield achieves a localization F1 score of 86.4% and invalidates 99% poisoned samples. Codes are released at https://github.com/Robin-WZQ/T2IShield.