Selective Learning Confusion Class for Text-Based CAPTCHA Recognition

Completely Automated Public Turing Test to Tell Computers and Humans Apart (CAPTCHA) is an important human-machine distinction technology for website to prevent the automatic malicious program attack. CAPTCHA recognition studies can find security breaches in CAPTCHA, improve CAPTCHA technology, it can also promote the technologies of license plate recognition and handwriting recognition. This paper proposed a method based on Convolutional Neural Network (CNN) model to identify CAPTCHA and avoid the traditional image processing technology such as location and segmentation. The adaptive learning rate is introduced to accelerate the convergence rate of the model, and the problem of over-fitting and local optimal solution has been solved. The multi task joint training model is used to improve the accuracy and generalization ability of model recognition. The experimental results show that the model has a good recognition effect on CAPTCHA with background noise and character adhesion distortion.

CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) systems serve as a crucial defense mechanism against automated attacks by distinguishing between human users and bots. However, advancements in deep learning have posed significant challenges to the security of conventional CAPTCHA systems. In this research paper, we present a comparative analysis of two prominent object detection algorithms, Convolutional Neural Networks (CNN) and Region-based Convolutional Neural Networks (RCNN), for breaking CAPTCHAs. The study evaluates the performance of CNN and RCNN algorithms in accurately identifying and deciphering characters within CAPTCHA images. Utilizing a diverse dataset of CAPTCHA samples, experiments are conducted to assess the effectiveness of both algorithms in handling variations in CAPTCHA styles, languages, and complexities. Through extensive experimentation and evaluation, we analyze the strengths and limitations of CNN and RCNN in the context of CAPTCHA breaking. Key metrics such as accuracy, precision, recall, and computational efficiency are compared to provide insights into the relative performance of each algorithm. The findings of this research contribute to the understanding of object detection techniques for CAPTCHA breaking and provide valuable insights for enhancing the security of CAPTCHA systems against emerging threats posed by deep learning-based attacks.

Multimedia Internet of Things (IoT) devices, especially, the smartphones are embedded with sensors including Global Positioning System (GPS), barometer, microphone, accelerometer, etc. These sensors working together, present a fairly complete picture of the citizens' daily activities, with implications for their privacy. With the internet, Citizens in Smart Cities are able to perform their daily life activities online with their connected electronic devices. But, unfortunately, computer hackers tend to write automated malicious applications to attack websites on which these citizens perform their activities. These security threats sometime put their private information at risk. In order to prevent these security threats on websites, Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHAs) are generated, as a form of security mechanism to protect the citizens' private information. But with the advancement of deep learning, text-based CAPTCHAs can sometimes be vulnerable. As a result, it is essential to conduct performance evaluation on the CAPTCHAs that are generated before they are deployed on multimedia web applications. Therefore, this work proposed an optimal and light-weight Convolutional Neural Network (CNN) to solve both numerical and alpha-numerical complex text-based CAPTCHAs simultaneously. The accuracy of the proposed CNN model has been accelerated based on Cyclical Learning Rates (CLRs) policy. The proposed CLR-CNN model achieved a high accuracy to solve both numerical and alpha-numerical text-based CAPTCHAs of 99.87% and 99.66%, respectively. In real-time, we observed that the speed of the model has increased, the model is lightweight, stable, and flexible as compared to other CAPTCHA solving techniques. The result of this current work will increase awareness and will assist multimedia security Researchers to continue and develop more robust text-based CAPTCHAs with their security mechanisms capable of protecting the private information of citizens in Smart Cities.

Online platforms' integrity and security are seriously threatened by the growth of automated bot attacks, necessitating the development of effective methods for telling harmful bots apart from legitimate users. In order to successfully combat automated bot attacks, this project investigates the creation and application of CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) methods utilising deep learning techniques. In this study, we employ deep learning to create and apply complex CAPTCHA-problems that help us distinguish between real users and automated bots. Convolutional neural networks and recurrent neural networks are employed to build dynamic, adjustable CAPTCHAs in order to keep up with the bots' evolving approaches. Our research focuses on creating CAPTCHAs that are simple to use for humans but difficult to understand by robots, which creates a favourable user experience for those who are actually human. Character skewing, background noise injection, and image obfuscation are some of the methods we use to safeguard our CAPTCHAs while maintaining their usability. Furthermore, we carry out exhaustive trials in the real world to assess the effectiveness of our methods based on in-depth CAPTCHA learning. We evaluate their resistance to a variety of attack methods, such as counterattacks and machine learning-based bot attacks, in order to confirm that they are resilient. The findings of our study show that utilising deep learning in CAPTCHA methods to thwart automated bot attacks is both feasible and effective. Our method makes the internet environment safer and more user-friendly while also enhancing security and reducing user annoyance. This work is an important step in the fight against the growing menace of automated bot attacks in the digital sphere.

CAPTCHAs (Completely Automated Public Turing test to tell Computers and Humans Apart) are widespread security measures on the World Wide Web that prevent automated programs from abusing online services. CAPTCHAs are designed to be easy for humans but hard for machines. Academic research into CAPTCHAs takes the form of a friendly ‘arms race’, with some researchers acting as ‘malicious users’ that try to attack and defeat the latest CAPTCHA systems automatically. Defeating a CAPTCHA requires two procedures: segmentation and recognition. Recent research shows that the problem of segmentation is much harder than recognition. The efforts to break the CAPTCHA determine its strength & thus in turn help the researcher to build stronger CAPTCHA. In order to test the strength of a particular CAPTCHA, in this paper, we have developed a universal algorithm that measures one of the parameters to measure strength of CAPTCHA. The proposed algorithm can adapt to various cases for segmenting characters from CAPTCHA image. The said algorithm is inspired from projection value of characters, concept of snake game and typical patterns of touching the characters. Experimental results prove that the proposed algorithm has improved correct segmentation rates giving accuracy of 85%.

In the area of image processing and Optical Character Recognition, Segmentation is one of the steps which plays an important role in dealing with offline and online text images. Character segmentation means breaking an image with word into a sequence of characters. A broad perspective of segmentation lies in segmenting the characters in CAPTCHA (Completely Automated Public Turing Tests to Tell Computers and Humans Apart). It is a test that authenticated users have to pass to gain access to their respective mail accounts. Malicious programs like bots, attack the accounts and are a threat to the data integrity, privacy and confidentiality. To avoid this, CAPTCHA was introduced. Segmentation of characters acts as the basis of analyzing the strength of CAPTCHA. Stronger the CAPTCHA, more difficult it is to break. The proposed work is about a CAPTCHA segmenter. It segments the CAPTCHA image with the help of a CAPTCHA Trainer. It is a user created set of pre-processing operations on images which can simply be re-used for segmenting the similar types of images, thus saving time. The operations that can be performed on any image are, gray scale conversion, dot removal, line removal, slant correction, color extraction, thinning.

The completely automated public Turing test to tell computers and humans apart (CAPTCHA) represents an established method to prevent automated abuse of web services. Most websites provide an audio CAPTCHA — in addition to a conventional visual scheme — to facilitate access for a wider range of users. These audio CAPTCHAs are generally based on distorted speech, rendering the task difficult for untrained or non-native listeners, while still being vulnerable against attacks that make use of automatic speech recognition techniques. In this work, we propose a novel and universally usable type of audio CAPTCHA that is solely based on the classification of acoustic sound events. We show that the proposed CAPTCHA leads to satisfactorily high human success rates, while being robust against recently proposed attacks, more than currently available speech-based CAPTCHAs.

CAPTCHAs (Completely Automated Public Turing test to tell Computers and Humans Apart) are increasingly used in many applications for machine and human identification. Compared with traditional English and digital characters based CAPTCHAs, Chinese characters contain more complicated characters which greatly enhance difficulty of automatic recognition. To solve that problem, we proposed a Convolution Neural Network (CNN) based approach. This approach greatly improves the recognition accuracy of Chinese Character CAPTCHAs with distortion, rotation and background noise. Our experiment results show that this approach achieves more than 95% accuracy for single character and 84% accuracy for three types of Chinese Character CAPTCHAs with four characters. This encouraging result indicates that deep neural network is useful in complicated structure perception of Chinese Character CAPTCHAs.

As the prevalence of online services and platforms continues to grow, so too does the threat posed by automated attacks seeking to circumvent security measures such as CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) mechanisms. This research paper delves into the various attack vectors and vulnerabilities inherent in CAPTCHA systems, aiming to provide a comprehensive understanding of the methodologies employed by attackers to bypass these safeguards. Through an investigation into machine learning-based algorithms, optical character recognition (OCR) techniques, and adversarial attacks, this study sheds light on the evolving landscape of CAPTCHA circumvention strategies. By analyzing the strengths and limitations of each attack vector, as well as the potential countermeasures available, this research contributes to the ongoing efforts to enhance the resilience of CAPTCHA systems against automated threats. Ultimately, the findings of this study serve to inform the development of more robust and adaptive CAPTCHA designs, thereby bolstering cyber security in the digital realm. This research paper focuses on investigating attack vectors and vulnerabilities in CAPTCHA mechanisms, including machine learning-based algorithms, OCR techniques, and adversarial attacks. It emphasizes the importance of understanding these methodologies to inform the development of more resilient CAPTCHA designs that can secure online platforms from various attacks.

The internet these days plays an important role in our daily lives, as the number of web services are increasing such as search engines and email, they are threatened by different types of attacks from computer programs such as robots. In order to solve this problem, CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart) was implemented to be used as a base to distinguish between human users and computer bots. Although this technique limits the automatic registration to web services to provide security but, still some CAPTCHAs have several weaknesses which allow the computer bots to infiltrate the mechanism of the CAPTCHA. The aim of this paper is to search for the problems in the existing types of CAPTCHA systems and to provide solutions or alternative approaches to solve these problem.

Internet is being used for several activities by a great range of users. These activities include communication, e-commerce, education, and entertainment. Users are required to register regarding Website in order to enroll Web activities. However, registration can be done by automated hacking software. That software make false enrollments which occupy the resources of the Website by reducing the performance and efficiency of servers, even stop the entire Web service. It is crucial for the Websites to have a system which has the capability of differing human users and computer programs in reading images of text. Completely automated public Turing test to tell computers and human apart (CAPTCHA) is such a defense system against optical character recognition (OCR) software. OCR can be defined as a software which work for defeating CAPTCHA images and make countless number of registration on the Websites. This study focuses on a new method which is splitting CAPTCHA images into several parts with random rotation values, and drawing random lines on a grid background. Lines are in the same color with the CAPTCHA text and they provide a distortion of image with grid background. In this paper, the algorithm of our method is introduced in detail.

Cloud computing is simply a metaphor for the internet. User does not requiredknowledge, control, and ownership in the computer infrastructure.User simply access or rent the software and paying only for what they use. Advantage of cloud computing is huge like Broad network access, Cost effectiveness, Rapid elasticity, Measured services, On -Demand service, Resource pooling , Location independence, Reliability, Energy saving and so on. But its global phenomenon that everything in this world has advantage as well as disadvantage, cloud computing also suffering from some drawback like security & privacy, Internet Dependency, Availability, And Current Enterprise Applications Can't Be Migrated Easily. I conclude that security is biggest hurdle in wide acceptance of cloud computing. U ser of cloud services are in fear of data loss, security and a vailability issues. At virtual level DDOS (Distributed Denial of Service Attack)is biggest threat of availability in cloud computing. In Denial of service attack an attacker prevent legitimate users of service from using the desired resources by flood a network or by consuming bandwidth .So authentication is need to distinguish legitimated clients from malicious clien ts, which can be performed through strong cryptographic verification (for a private server) or graphical Turing tests (for a public server). Where the authentication is performed by Graphical Turing Tests, which is widely used to distinguish human users from robots through their reaction. On the other hand, CAPTCHA (Completely Automated Public Turing Tests to Tell Computers and Humans Apart) is used for Graphical Turing Test. There are many OCR or Non -OCR based CAPTCHA’s are used widely but they are vulnerable to many attacks likePixel-Count Attack, Recognition by using OCR, Dictionary Attack, and Vertical Segmentation. This paper introduces a new CAPTCHA method called Two-Tier CAPTCHA.In this method CLAD node need to generate two things,first a alphanumeric CAPTCHA code with image. Second Query related to that CAPTCHA code. E.g. enter only Digit’s .We can increase the rate of its difficulty in order to improve its resistance against the attacks by ad ding more and more query and combination in database.The algorithm of this method makes it hard for bot programs which mean that it is more secure.This project has been implemented byASP.NET and PHP Language.

With the prosperity of SMS (short message service), there are more and more SMS spam. This paper proposes a method for filtering group sending SMS Spam, which combines black-and-white list technology and CAPTCHA (completely automated public Turing test to tell computers and human apart) technique. At first, it utilizes black-and-white list technique to check out the SMS account. If the SMS account is on the blacklist or white list, the SMS will be disposed directly without using CAPTCHA technique. In contrast, if the SMS account is not on the black-and-white list, then it makes use of CAPTCHA to handle it. If the SMS sender can not pass the CAPTCHA, the SMS will be tagged as s SMS spam and deleted. Finally, it has been simulated with Smartphone of Microsoft Visual Studio 2005.

As of the advanced Completely Automated Public Turing tests to tell Computers and Humans Apart (CAPTCHAs), the CAPTCHA using mental rotation has been proposed. Mental rotation is an advanced human-cognitive-processing ability to rotate mental representations of one single 2D/3D object. However, as have already been reported, the mental rotation CAPTCHA can be overcome by pattern matching and/or machine learning. Therefore, this paper proposes to enhance the mental rotation CAPTCHA by using two distinct 3D objects in the task of mental rotation, which we call mental rotation. We implemented a prototype of the sophisticated mental rotation CAPTCHA, and carried out basic experiments to confirm its usability. Also, we conducted a comparison between the proposed CAPTCHA and existing CAPTCHAs. The obtained results were satisfactory.

There are many websites specially designed for mobile phones. Some hackers write automated programs to abuse these website services and waste the website resources. Therefore, it is necessary to distinguish between human users and computer programs. Methods known for achieving this are known as CAPTCHA (Completely Automated Public Turing test to tell Computers and Human Apart). CAPTCHA methods are mainly based on the weaknesses of OCR (Optical Character Recognition) systems and ask the user to type a word. So using them is difficult in tools such as PDAs (Personal Digital Assistant) or mobile phones that lack a complete keyboard. In this paper, a new CAPTCHA system is proposed for touch-screen devices such as PDAs and mobile phones. In this system, a word is drawn in a random place on the screen and a number of arcs are drawn on the screen. Then the user is asked to highlight the word by the stylus. Due to the limitations of the PDA and mobile phone, OCR programs on these devices cannot recognize the shown word, while a human user can easily highlight the word. The proposed method is implemented by the JavaME (Java Platform Micro Edition) language and tested on a Sony Ericsson P990i mobile phone.