fgCAPTCHA combines images of human faces, cartoons, and sketches on a multi-colored background. To prevent software attacks on the CAPTCHA, two visual distortions are applied to each CAPTCHA using parameters optimized by a genetic algorithm. This process ensures the best possible balance between ease-of-use for humans and resistance to machine attacks.
To solve fgCAPTCHA, users must tap or click on all the human face photographs.
Results
In testing, humans were able to successfully solve 87.9% of fgCAPTCHA images with some distortions have success rates of up to 97%. As part of the generation process, CAPTCHAs solved by face detection are automatically removed yielding a machine attack rate of approximately 0%.
Demos
Publications
Journal Articles
- B. M. Powell, G. Goswami, M. Vatsa, R. Singh, and A. Noore, “fgCAPTCHA: Genetically Optimized Face Image CAPTCHA,” IEEE Access, vol. PP, no. 99, Apr. 2014. PDF Publisher's Website