Web services accounts have recently been automatically acquired in large quantities by bot-programs, which are malicious. Furthermore, the acquired accounts have been used for spamming, which is a problem for service operators or Internet users. Completely Automated Public Turing Tests To Tell Computers and Humans Apart (CAPTCHAs) have generally been adopted for Web Services as a method of preventing Web services accounts from being acquired. These are Turing tests for users of Web services to distinguish between humans and bot-programs. There are several types of methods in CAPTCHAs, but the most typical in this field are text-based. However, these methods can be decoded with a high degree of probability, because of the OCR technology that has evolved. Much research on resolving this problem has been proposed. For example, one method in this research adds distortion to characters to make it difficult to analyze them with OCR. Despite this research, OCR has a higher success rate than humans. Thus, we propose a practical method for CAPTCHA in this paper in which only humans can provide correct answers by applying a modal completion.
展开▼
