We always use CAPTCHA(Completely Automated Public Turing test to Tell Computers and Humans Apart) to prevent automated bot for data entry. Although there are various kinds of CAPTCHAs, text-based scheme is still applied most widely, because it is one of the most convenient and user-friendly way for daily user [1]. The fact is that segmentations of different types of CAPTCHAs are not always the same, which means one of CAPTCHA's bottleneck is the segmentation. Once we could accurately split the character, the problem could be solved much easier. Unfortunately, the best way to divide them is still case by case, which is to say there is no universal way to achieve it. In this paper, we present a novel algorithm to achieve state-of-the-art performance, what was more, we also constructed a new convolutional neural network as an add-on recognition part to stabilize our state-of-the-art performance of the whole CAPTCHA system. The CAPTCHA datasets we are using is from the State Administration for Industry& Commerce of the People's Republic of China. In this datasets, there are totally 33 entrances of CAPTCHAs. In this experiments, we assume that each of the entrance is known. Results are provided showing how our algorithms work well towards these CAPTCHAs.

Combining convolutional neural network and self-adaptive algorithm to defeat synthetic multi-digit text-based CAPTCHA

A survey of CAPTCHA technologies to distinguish between human and computer

CAPTCHA (Completely Automated Public Turing test to Tell Computers and Humans Apart) is almost everywhere in data entry due to automated scripts like bots. Nowadays, text-based scheme is still applied most widely, which typically need the users to answer questions regarding recognition task. In particular, the segmentations of different types of CAPTCHAs are not always the same. As so far, there isn't any universal way to solve the segmentation problems. In this paper, we present a novel adaptive algorithm and based on that we create a system to defeat several CAPTCHAs at the same time. The CAPTCHA datasets we used are from the State Administration for Industry&Commerce of the People's Republic of China. There are totally 33 entrances of CAPTCHAs we need to solve. In this experiments, we assume that each of the entrance is known. Results are provided showing how our algorithms work well towards these CAPTCHAs.

A self-adaptive algorithm to defeat text-based CAPTCHA

Introduction In the era of Cyber threats, it is very important to understand if the user trying to access a Website is a real person or a "bot", a malicious automated program. In order to decide whether to allow the access, the first CAPTCHA test was invented in 2000 by John Langford, Nicholas J. Hooper and Luis Von Ahn and it is still used (Singh & Pal, 2014). CAPTCHA is an acronym for "Completely Automated Public Turing test to tell Computers and Humans Apart". The CAPTCHA test performs an authentication process, called a "challenge-response authentication", because it presents a challenge to the user, and only when it is solved, the right to access the Website is given (Ling-Zi & Yi-Chun, 2012). It is difficult and almost impossible for bots to solve these challenges (Saini & Bala, 2013). The operation is similar to a reverse "Turing Test" where the CAPTCHA test acts like a judge and on the other side stands the "user", which may be human or not. CAPTCHA tests must be, on the one hand, very easy for the user in order to pass, and, on the other hand, very difficult for the bots to pass. The conflict between approachability and protection in the Internet world is a complex issue, dealing with the necessity to allow a wide range of different users to access the desired Website, but preventing the access of malicious elements. Cyber threats became abundant, and the attempts to reach computers by unauthorized agents are growing. For example, according to Symantec (2011), 75.8% of the emails sent are "spam", where 82.2% of them are originated by bots. Therefore, CAPTCHA implementations can be found on more than 3.5 million sites globally, in all kind of Websites, like sites for fill-in forms, write comments, and buy tickets, and human beings solve CAPTCHA tests more than 300 million times a day (Angre, Kapadia, & Ugale, 2015). Different Types of CAPTCHA Tests During the last years, several types of CAPTCHA tests were defined and developed. Below is a description of the most common different types of tests. Text-based CAPTCHA tests--This is the most common test, called "reCAPTCHA", in which a sequence of letters and/or numbers is shown in a distorted and twisted manner, as seen in Figure 1. The user needs to identify and decipher what is shown, and type the exact sequence into a provided text-box in order to get access to the desired Webpage. This system, was originally aimed to help digitize printed text that was hard to read by OCR (Optical Character Recognition) softwares, and was acquired by Google in 2009, in order to use the "Common Wisdom" to digitize large number of manuscripts ("reCAPTCHA," n.d.). Audio based CAPTCHA tests--This test was developed for vision-impaired people. The user must listen to a sequence of letters and/or numbers and type the sequence in the text-box. Pre-analysis of user behavior--In 2013, reCAPTCHA began implementing behavioral analysis of the browser's interactions with the CAPTCHA to predict whether the user is a human or a bot. This analysis, called "No CAPTCHA reCAPTCHA" (shown in Figure 2), occurs before displaying the text-based CAPTCHA and, according to the analysis, when there are reasons to think the user might be a bot, a more difficult test is presented to the user. By end of 2014 this mechanism started to be rolled out to most of the public Google services ("reCAPTCHA," n.d.). Arithmetic operation based CAPTCHA tests--In this type of test, as shown in Figure 3, the sequence shown to the user contains a very basic arithmetic operation, i.e., "1 + 3=", which can be performed by almost every human being. The user needs to enter the result of the operation into a text box. Picture based CAPTCHA tests--In this test a number of pictures are shown to the user, with a simple question, as shown in Figure 4. The user must identify the pictures and select the one that is the correct answer to the question. …

/pdf/captcha-security-affecting-user-experience-58xjwv0ynb.pdf

CAPTCHA – Security affecting User Experience

CAPTCHA(Completely Automated Public Turing test to Tell Computers and Humans Apart) can be used to protect data from auto bots. Countless kinds of CAPTCHAs are thus designed, while we most frequently utilize text-based scheme because of most convenience and user-friendly way \cite{bursztein2011text}. Currently, various types of CAPTCHAs need corresponding segmentation to identify single character due to the numerous different segmentation ways. Our goal is to defeat the CAPTCHA, thus firstly the CAPTCHAs need to be split into character by character. There isn't a regular segmentation algorithm to obtain the divided characters in all kinds of examples, which means that we have to treat the segmentation individually. In this paper, we build a whole system to defeat the CAPTCHAs as well as achieve state-of-the-art performance. In detail, we present our self-adaptive algorithm to segment different kinds of characters optimally, and then utilize both the existing methods and our own constructed convolutional neural network as an extra classifier. Results are provided showing how our system work well towards defeating these CAPTCHAs.

/pdf/an-optimized-system-to-solve-text-based-captcha-57cba4w6bi.pdf

An optimized system to solve text-based CAPTCHA

CAPTCHA (Completely Automated Public Turing Test to tell Computers and Human Apart) is widely used than before, which becomes the common part of current website login system. However, the CAPTCHA implementation is tricky and risky without deliberate design. In this paper, we give a study case of the vulnerabilities in current login website using text-based CAPTCHA. Our target is a website of mainstream bank of china. We show that with some specialized methods, the CAPTCHA scheme in its website can be easily cracked. Finally, we give some advices for CAPTCHA designers to revise our CAPTCHA implementation security in the future.

A Case Study of Text-Based CAPTCHA Attacks

Many researches prove that MIMO can enlarge the channel capacity without extra bandwidth and transmission power. This paper uses the MIMO technology in mobile multimedia system, describes the increasing of the channel capacity. Then, this paper describes the application of CSIT and calculates the capacity gain caused by it. At last, the benefit of the CSIT is proved by computer simulations on the base of DAB/DMB standard.

The capacity gain of mobile multimedia broadcasting system by using CSIT in MIMO

This paper introduced a novel method for implementing broadcast encryption. Our scheme takes advantages of bilinear map and group characteristic, and shifts most of the storage overhead to the public device instead of storing in the tamper-proof device which is a major problem on current implementation. Furthermore, the broadcast keys in our scheme could be reused periodically resulting in more operational efficiency.

Zhang Yichun

Papers

A Case Study of Text-Based CAPTCHA Attacks

The capacity gain of mobile multimedia broadcasting system by using CSIT in MIMO

Implementing broadcast encryption scheme using bilinear map and group characteristic