Research On Chinese Audio CAPTCHA Technique And Application

As one of the most popular techniques in the field of computer security and human-computer interaction, CAPTCHA(Completely Automated Public Turing Test to Tell Computers and Humans Apart) is widely used in the Internet field. Most of the CAPTCHA techniques that has been involved in the daily life is visual-based. However, this type of visual-based CAPTCHA is considerably inconvenient for the visual impaired people. Currently there are only limited amount of websites which take the audio CAPTCHA as the general application. Moreover, most of the available applications are based on English. The aforementioned facts results in significant inconvenience who speak as foreign language.As a special group in the society, the career for disable people is an important part of China's socialist modernization. Accessibility is one of the key points to ensure that disable people can be involved into the social life. Information accessibility as the kernel of accessibility environment deserves more attention and care. As cross point of Internet technology and information accessibility, a new concept of CAPTCHA needs to be developed for both disable and healthy people.In this dissertation, the design and development of Chinese Audio CAPTCHA were proposed, and two important parameters of CAPTCHA, namely the robust and the practicability, were analyzed. The pure audio files were collected and the tones of audio were modified. After that, the files were mixed with selected kinds of noises, generating the Chinese Audio CAPTCHA. Three classifier algorithms were employed to classify the Chinese Audio CAPTCHA. The results were also compared with outcome obtained from experimenters, to obtain the methodology that was able to generate the Audio CAPTCHA with the most enhanced robustness and practicability.To sum up, the key contributions in this dissertation are listed below:(1) Proposing the technique of adding noise in to Chinese Audio CAPTCHA, which could used by both Chinese visual-disabled people and health people.(2) Developing the technique of Chinese Audio CAPTCHA which based on the live voice library.(3) Proposing the technique of evaluation of Chinese Audio CAPTCHA by multi-classifiers and manual classifier. Evaluate the Chinese Audio CAPTCHA, and analysis the result data to proof the best method to generate Chinese Audio CAPTCHA.(4) Proposing the method of deploying the research result on WebService platform.