Intelligent Graphical Authentication System with Adaptive Security, AI-Driven Enhancements and Behavioral Analysis

This paper presents an Intelligent Graphical Authentication System that integrates multiple layers of security to enhance both usability and protection against unauthorized access. The system combines three key authentication steps: image selection, a grid-based password with dynamic strength evaluation, and Time-Based One-Time Password (TOTP) verification. Users first select four images from a designated domain during registration, which must be reselected in the same order during login. The grid-based password selection requires users to choose points from a 7x7 grid with strength penalties for adjacent or diagonal points, guided by an AI-driven Markov Model that suggests secure point combinations. Additionally, TOTP adds an extra layer of protection by generating time-sensitive codes for login. The system incorporates behavioral analysis to monitor login patterns, blocking accounts after multiple failed attempts and enabling recovery via OTP and TOTP.