The ability of hackers to infiltrate computer systems using computer attack programs and bots led to the development of Captchas or Completely Automated Public Turing Tests to Tell Computers and Humans Apart. The text Captcha is the most popular Captcha scheme given its ease of construction and user friendliness. However, the next generation of hackers and programmers has decreased the expected security of these mechanisms, leaving websites open to attack. Text Captchas are still widely used, because it is believed that the attack speeds are slow, typically two to five seconds per image, and this is not seen as a critical threat. In this paper, we introduce a simple, generic, and fast attack on text Captchas that effectively challenges that supposition. With deep learning techniques, our attack demonstrates a high success rate in breaking the Roman-character-based text Captchas deployed by the top 50 most popular international websites and three Chinese Captchas that use a larger character set. These targeted schemes cover almost all existing resistance mechanisms, demonstrating that our attack techniques are also applicable to other existing Captchas. Does this work then spell the beginning of the end for text-based Captcha? We believe so. A novel image-based Captcha named Style Area Captcha (SACaptcha) is proposed in this paper, which is based on semantic information understanding, pixel-level segmentation, and deep learning techniques. Having demonstrated that text Captchas are no longer secure, we hope that our proposal shows promise in the development of image-based Captchas using deep learning techniques.
1 July 2021 in IEEE Transactions on Image Processing
Convolutional neural networks (CNNs) are vulnerable to being deceived by adversarial examples generated by adding small, human-imperceptible perturbations to a clean image. In this paper, we propose an image reconstruction network that reconstructs an input adversarial example into a clean output image to defend against such adversarial attacks. Due to the powerful learning capabilities of the residual block structure, our model can learn a precise mapping from adversarial examples to reconstructed examples. The use of a perceptual loss greatly suppresses the error amplification effect and improves the performance of our reconstruction network. In addition, by adding randomization layers to the end of the network, the effects of additional noise are further suppressed, especially for iterative attacks. Our model has the following four advantages. 1) It greatly reduces the impact of adversarial perturbations while having little influence on the prediction performance of clean images. 2) During inference phase, it performs better than most existing model-agnostic defense methods. 3) It has better generalization capability. 4) It can be flexibly combined with other methods, such as adversarially trained models.
15 July 2019 in IEEE Transactions on Information Forensics and Security
Text-based CAPTCHAs are the most widely used CAPTCHA scheme. Most text-based CAPTCHAs have been cracked. However, previous works have mostly relied on a series of preprocessing steps to attack text CAPTCHAs, which was complicated and inefficient. In this paper, we introduce a simple, generic and effective end-to-end attack on text CAPTCHAs without any preprocessing. Through a convolutional neural network and an attention-based recurrent neural network, our attack broke a wide range of real-world text CAPTCHAs that are deployed by the top 50 most popular websites ranked by Alexa.com. Additionally, this paper comprehensively analyzed the security of most resistance mechanisms of text-based CAPTCHAs through experiments. The experimental results prove that the anti-segmentation principle can be completely broken under deep learning attacks without any segmentation or preprocessing steps, in contrast to commonly held beliefs.
15 March 2017 in IEEE Transactions on Information Forensics and Security
Captcha is a security mechanism designed to differentiate between computers and humans, and is used to defend against malicious bot programs. Text-based Captchas are the most widely deployed differentiation mechanism, and almost all text-based Captchas are single layered. Numerous successful attacks on the single-layer text-based Captchas deployed by Google, Yahoo!, and Amazon have been reported. In 2015, Microsoft deployed a new two-layer Captcha scheme. This appears to be the first application of two-layer Captchas. It is, therefore, natural to ask a fundamental question: is the two-layer Captcha as secure as its designers expected? Intrigued by this question, we have for the first time systematically analyzed the security of the two-layer Captcha in this paper. We propose a simple but an effective method to attack the two-layer Captcha deployed by Microsoft, and achieve a success rate of 44.6% with an average speed of 9.05 s on a standard desktop computer (with a 3.3-GHz Intel Core i3 CPU and 2-GB RAM), thus demonstrating clear security issues. We also discuss the originality and applicability of our attack, and offer guidelines for designing Captchas with better security and usability.
Text-based CAPTCHA has become one of the most popular methods for preventing bot attacks. With the rapid development of deep learning techniques, many new methods to break text-based CAPTCHAs have been developed in recent years. However, a holistic and uniform investigation and comparison of these attacks’ effects is lacking due to inconsistent choices of model structures, training datasets, and evaluation metrics. In this article, we perform an experimental investigation on the effects of existing attacks on text-based CAPTCHA schemes. We first summarize existing text-based CAPTCHAs using a newly proposed taxonomy based on their resistance mechanisms and systematically review corresponding attacks in terms of methods and pros/cons. Then, we introduce a unified attack framework that contains a number of different attack modules and transfer learning strategies. Applying this framework, we extensively evaluate the performance of known attacks on 20 CAPTCHA schemes in terms of accuracy and efficiency; then, we investigate the robustness of these widely used schemes and discover the effects of previously unexplored attacks. Finally, we discuss future CAPTCHA designs based on our experimental results and findings. Our work also contributes to the CAPTCHA community by offering an open-access dataset that contains 22 different CAPTCHA sample sets.
CAPTCHA, or Completely Automated Public Turing Tests to Tell Computers and Humans Apart, is a common mechanism used to protect commercial accounts from malicious computer bots, and the most widely used scheme is text-based CAPTCHA. In recent years, newly emerged deep learning techniques have achieved high accuracy and speed in attacking text-based CAPTCHAs. However, most of the existing attacks have various disadvantages, the attack process made high complexity or manually collecting and labeling a large number of samples to train a deep learning recognition model is time-consuming and expensive. In this paper, we propose a transfer learning-based approach that greatly reduces the attack complexity and the cost of labeling samples, specifically, by pre-training the model with randomly generated samples and fine-tuning the pre-trained model with a small number of real-world samples. To evaluate our attack, we tested 25 online CAPTCHAs achieving success rates ranging from 36.3% to 96.9%. To further explore the effect of the training sample characteristics on the attack accuracy, we elaborately imitate some samples and apply a generative adversarial network to refine the samples, sequentially we use these two kinds of generated samples to pre-train the models, respectively. The experimental results demonstrate that the similarity between randomly generated samples and elaborately imitated samples has a negligible impact on the attack accuracy. Instead, transfer learning is the key factor; it reduces the cost of data preparation while preserving the model’s attack accuracy.