Automated CAPTCHA Generation Using Machine Learning For Image And Audio Challenges

Mohammed Abdul Rahman; Maher Nawaz Qadri; Mohammed Furqan Ahmed; Dr. Abdul Ahad Afroz

doi:10.62647/IJITCE2025V13I2sPP253-261

Authors

Mohammed Abdul Rahman 1,2,3B.E Students, Department of Information Technology, ISL Engineering College, Hyderabad, India. Author
Maher Nawaz Qadri B.E Students, Department of Information Technology, ISL Engineering College, Hyderabad, India. Author
Mohammed Furqan Ahmed B.E Students, Department of Information Technology, ISL Engineering College, Hyderabad, India. Author
Dr. Abdul Ahad Afroz Associate Professor, Department of Information Technology, ISL Engineering College, Hyderabad, India. Author

DOI:

https://doi.org/10.62647/IJITCE2025V13I2sPP253-261

Keywords:

CAPTCHA Systems, Accessibility, Usability, Audio CAPTCHA, Text-based CAPTCHA, Human-Computer Interaction, Adaptive Security, Bot Prevention

Abstract

Advancements in CAPTCHA design are reshaping digital security and accessibility. This study introduces an adaptive system that evaluates audio, text-based, and emerging CAPTCHA formats to balance usability and protection against bots. A key focus is SoundsRight, an audio CAPTCHA aiding visually impaired users by using sound recognition amidst noise. User trials revealed improved security but longer response times, highlighting usability trade-offs. Text-based CAPTCHAs tested on tablets showed performance variations based on demographics like age and experience. Additional formats—image, gesture, logic, and gamified—were assessed for inclusivity and effectiveness. Results support hybrid CAPTCHA systems that adapt to diverse user needs, ensuring broader accessibility without compromising security