Segmentation-validation based handwritten Arabic CAPTCHA generation

摘要

A Completely Automated Public Turing test to tell Computers and Human Apart (CAPTCHA) is an effective way to identify human users and prevent cyber-attacks. In this research, we present a novel approach for generating and validating handwritten Arabic CAPTCHAs. With high Internet penetration among Arabic speakers, Arabic CAPTCHAs have recently gained interests among the researchers. Handwritten text based CAPTCHAs have been shown as more secured compared to printed text based CAPTCHAs. Therefore, the proposed method uses synthesized handwritten Arabic words to generate CAPTCHAs. In the user-validation phase of our method, users solve the CAPTCHAs by finding the segmentation points in the cursive Arabic words. To the best of our knowledge, such a segmentation-validation based approach for CAPTCHAs is reported for the first time in this paper. In addition, this work introduces the use of synthesized words for handwritten Arabic CAPTCHAs. The proposed method is sufficiently flexible to generate CAPTCHAs with various degrees of difficulty. The usability and security aspects of the current method have been evaluated through experimental studies with very promising results.