MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202621057437 A) filed by Uplenchwar Riddhi Pravin; and Deshmukh Vishwesh on May 06, 2026, for A System And Method For Continuous Authentication Using Behavioral Biometrics For Fraud Detection..
Inventors include Uplenchwar Riddhi Pravin; Umap Vishakha Pradeep; Utture Anand Dinesh; Vadher Manasvi Rajendrasinh; Uphade Siddhesh Vishwanath; and Deshmukh Vishwesh.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: The present invention relates to a computer-implemented system and method for continuous authentication and fraud detection using behavioral fingerprinting. The system captures multi-modal user interaction data including keystroke dynamics, mouse movement patterns, touch interactions, and device attributes through a client- side module. A preprocessing module generates structured behavioral features, which are used to establish and continuously update a personalized behavioral baseline comprising a rolling baseline and a reference baseline. A dynamic multi-modal fusion engine computes a behavioral risk score representing deviation from the baseline using weights assigned based on signal reliability, consistency, and contextual relevance. The system further incorporates a honeypot interaction module that embeds dynamically generated decoy interface elements to detect automated or synthetic interactions. Honeypot-derived anomaly signals are integrated into the risk scoring mechanism. Machine learning models classify user behavior, and a confidence score is generated based on model agreement and data quality. A decision engine dynamically controls system access states, enabling continuous, adaptive, and real-time fraud detection.
Disclaimer: Curated by HT Syndication.
