When Vision Fails: Text Attacks Against ViT and OCR

• URL: http://arxiv.org/abs/2306.07033v2
• Date: Sat, 11 Oct 2025 04:55:43 GMT
• Title: When Vision Fails: Text Attacks Against ViT and OCR
• Authors: Nicholas Boucher, Jenny Blessing, Ilia Shumailov, Ross Anderson, Nicolas Papernot,
• Abstract summary: Text-based machine learning models are vulnerable to an emerging class of Unicode-based adversarial examples.<n>In theory, OCR models will ignore any malicious Unicode characters and will extract the visually correct input to be fed to the model.<n>We show that these visual defenses fail to prevent this type of attack.
• Score: 37.010684530076205
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Text-based machine learning models are vulnerable to an emerging class of Unicode-based adversarial examples capable of tricking a model into misreading text with potentially disastrous effects. The primary existing defense against these attacks is to preprocess potentially malicious text inputs using optical character recognition (OCR). In theory, OCR models will ignore any malicious Unicode characters and will extract the visually correct input to be fed to the model. In this work, we show that these visual defenses fail to prevent this type of attack. We use a genetic algorithm to generate visual adversarial examples (i.e., OCR outputs) in a black-box setting, demonstrating a highly effective novel attack that substantially reduces the accuracy of OCR and other visual models. Specifically, we use the Unicode functionality of combining characters (e.g., \~n which combines the characters n and ~) to manipulate text inputs so that small visual perturbations appear when the text is displayed. We demonstrate the effectiveness of these attacks in the real world by creating adversarial examples against production models published by Meta, Microsoft, IBM, and Google. We additionally conduct a user study to establish that the model-fooling adversarial examples do not affect human comprehension of the text, showing that language models are uniquely vulnerable to this type of text attack.

Related papers

• TextInVision: Text and Prompt Complexity Driven Visual Text Generation Benchmark [61.412934963260724]
  Existing diffusion-based text-to-image models often struggle to accurately embed text within images.<n>We introduce TextInVision, a large-scale, text and prompt complexity driven benchmark to evaluate the ability of diffusion models to integrate visual text into images.
  arXiv  Detail & Related papers  (2025-03-17T21:36:31Z)
• Web Artifact Attacks Disrupt Vision Language Models [61.59021920232986]
  Vision-language models (VLMs) are trained on large-scale, lightly curated web datasets. They learn unintended correlations between semantic concepts and unrelated visual signals. Prior work has weaponized these correlations as an attack vector to manipulate model predictions. We introduce artifact-based attacks: a novel class of manipulations that mislead models using both non-matching text and graphical elements.
  arXiv  Detail & Related papers  (2025-03-17T18:59:29Z)
• Vision-LLMs Can Fool Themselves with Self-Generated Typographic Attacks [58.10730906004818]
  Typographic attacks, adding misleading text to images, can deceive vision-language models (LVLMs)<n>Our experiments show these attacks significantly reduce classification performance by up to 60%.
  arXiv  Detail & Related papers  (2024-02-01T14:41:20Z)
• SA-Attack: Improving Adversarial Transferability of Vision-Language Pre-training Models via Self-Augmentation [56.622250514119294]
  In contrast to white-box adversarial attacks, transfer attacks are more reflective of real-world scenarios. We propose a self-augment-based transfer attack method, termed SA-Attack.
  arXiv  Detail & Related papers  (2023-12-08T09:08:50Z)
• Invisible Threats: Backdoor Attack in OCR Systems [26.471281625129226]
  This work proposes a backdoor attack for OCR resulting in the injection of non-readable characters from malicious input images. This simple but effective attack exposes the state-of-the-art OCR weakness, making the extracted text correct to human eyes but simultaneously unusable for the NLP application.
  arXiv  Detail & Related papers  (2023-10-12T12:05:51Z)
• Are aligned neural networks adversarially aligned? [93.91072860401856]
  adversarial users can construct inputs which circumvent attempts at alignment. We show that existing NLP-based optimization attacks are insufficiently powerful to reliably attack aligned text models. We conjecture that improved NLP attacks may demonstrate this same level of adversarial control over text-only models.
  arXiv  Detail & Related papers  (2023-06-26T17:18:44Z)
• I See Dead People: Gray-Box Adversarial Attack on Image-To-Text Models [0.0]
  We present a gray-box adversarial attack on image-to-text, both untargeted and targeted. Our attack operates in a gray-box manner, requiring no knowledge about the decoder module. We also show that our attacks fool the popular open-source platform Hugging Face.
  arXiv  Detail & Related papers  (2023-06-13T07:35:28Z)
• Verifying the Robustness of Automatic Credibility Assessment [50.55687778699995]
  We show that meaning-preserving changes in input text can mislead the models. We also introduce BODEGA: a benchmark for testing both victim models and attack methods on misinformation detection tasks. Our experimental results show that modern large language models are often more vulnerable to attacks than previous, smaller solutions.
  arXiv  Detail & Related papers  (2023-03-14T16:11:47Z)
• Character-Aware Models Improve Visual Text Rendering [57.19915686282047]
  Current image generation models struggle to reliably produce well-formed visual text. Character-aware models provide large gains on a novel spelling task. Our models set a much higher state-of-the-art on visual spelling, with 30+ point accuracy gains over competitors on rare words.
  arXiv  Detail & Related papers  (2022-12-20T18:59:23Z)
• Vision-Language Pre-Training for Boosting Scene Text Detectors [57.08046351495244]
  We specifically adapt vision-language joint learning for scene text detection. We propose to learn contextualized, joint representations through vision-language pre-training. The pre-trained model is able to produce more informative representations with richer semantics.
  arXiv  Detail & Related papers  (2022-04-29T03:53:54Z)
• Language Matters: A Weakly Supervised Pre-training Approach for Scene Text Detection and Spotting [69.77701325270047]
  This paper presents a weakly supervised pre-training method that can acquire effective scene text representations. Our network consists of an image encoder and a character-aware text encoder that extract visual and textual features. Experiments show that our pre-trained model improves F-score by +2.5% and +4.8% while transferring its weights to other text detection and spotting networks.
  arXiv  Detail & Related papers  (2022-03-08T08:10:45Z)
• Bad Characters: Imperceptible NLP Attacks [16.357959724298745]
  A class of adversarial examples can be used to attack text-based models in a black-box setting. We find that with a single imperceptible encoding injection an attacker can significantly reduce the performance of vulnerable models. Our attacks work against currently-deployed commercial systems, including those produced by Microsoft and Google.
  arXiv  Detail & Related papers  (2021-06-18T03:42:56Z)
• Detecting Adversarial Examples by Input Transformations, Defense Perturbations, and Voting [71.57324258813674]
  convolutional neural networks (CNNs) have proved to reach super-human performance in visual recognition tasks. CNNs can easily be fooled by adversarial examples, i.e., maliciously-crafted images that force the networks to predict an incorrect output. This paper extensively explores the detection of adversarial examples via image transformations and proposes a novel methodology.
  arXiv  Detail & Related papers  (2021-01-27T14:50:41Z)
• Adversarial Watermarking Transformer: Towards Tracing Text Provenance with Data Hiding [80.3811072650087]
  We study natural language watermarking as a defense to help better mark and trace the provenance of text. We introduce the Adversarial Watermarking Transformer (AWT) with a jointly trained encoder-decoder and adversarial training. AWT is the first end-to-end model to hide data in text by automatically learning -- without ground truth -- word substitutions along with their locations.
  arXiv  Detail & Related papers  (2020-09-07T11:01:24Z)
• Visual Attack and Defense on Text [18.513619521807286]
  Modifying characters of a piece of text to their visual similar ones often ap-pear in spam in order to fool inspection systems and other conditions. We ap-ply a vision-based model and adversarial training to defense the attack without losing the ability to understand normal text.
  arXiv  Detail & Related papers  (2020-08-07T15:44:58Z)
• BAE: BERT-based Adversarial Examples for Text Classification [9.188318506016898]
  We present BAE, a black box attack for generating adversarial examples using contextual perturbations from a BERT masked language model. We show that BAE performs a stronger attack, in addition to generating adversarial examples with improved grammaticality and semantic coherence as compared to prior work.
  arXiv  Detail & Related papers  (2020-04-04T16:25:48Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.