关键词: 验证码, 图像旋转, 模板匹配, 序贯相似性检测算法

Abstract:

The CAPTCHA is a Turing Test which is used to tell computers from humans. This paper presents the sequential similarity detection algorithm(SSDA) algorithm based on Rough-matching to crack the CAPTCHA with characters rotating and non-distorting. The characters are rotated to normalize the angle of rotation to reduce the template library, with the Rough-matching algorithm used to shorten the time consumption. Taking the CAPTCHA of a bank as an example, we document the four breaking parts, which are Preprocessing, Extraction, Rotation and Recognition. Experimental results show that the success rate of the SSDA matching algorithm based on coarse matching is 85％, and that the average breaking time is 3.5s. It is much better than a similar algorithm.

Key words: CAPTCHA, image rotation, template matching, sequential similarity detection algorithm