MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202621057010 A) filed by Dr. D. Y. Patil Institute Of Technology, Pimpri, Pune - on May 05, 2026, for Multi-Layer Real-Time Transaction Fraud Detection And Behavioral Biometric Session Monitoring System.
Inventors include Amaan Ahmed Shakeel Ahmed Shaikh; Ms. Shraddha Shingne; Mrs. Swati Khairnar; and Mrs. Priyanka Chetwani.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: This invention discloses an integrated, three-layer middleware architecture for real-time fraud detection and comprehensive security monitoring in high-throughput financial transaction environments. The first layer performs ultra-low-latency transaction evaluation by combining a deterministic rules engine, a gradient-boosted decision model (LightGBM) optimized through ONNX Runtime, and a configurable policy engine. This layer classifies each transaction into one of five predefined risk decision codes with an average processing latency of approximately 0.46 milliseconds, ensuring rapid and consistent decision-making. The second layer operates at the API gateway level, providing continuous surveillance of incoming and outgoing requests. It classifies activities into seven distinct threat categories and enforces a five-tier token-bucket rate-limiting mechanism to prevent abuse and denial-of-service patterns. High- severity threats are automatically blocked in real time, while suspicious or borderline cases are escalated to a Security Operations Center, supported by comprehensive audit trails for traceability and compliance. The third layer focuses on session-level behavioral analytics, continuously evaluating user interactions using five weighted anomaly detection signals. These signals are aggregated into a dynamic risk score ranging from 0 to 100. When the score exceeds a predefined threshold of 80, the system proactively terminates the session and can override prior transaction approvals to mitigate emerging risks. By consolidating functionalities typically distributed across multiple enterprise systems, the proposed middleware delivers a unified, scalable, and deployable solution that ensures explainability, auditability, and sub-millisecond decision performance for banking, payment, and insurance platforms.
Disclaimer: Curated by HT Syndication.
