MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202641070221 A) filed by Venkatarathnam Korukonda; Kadali Satyanarayana; Gunavardini V, Cse-Bit; Vikash Sawan, Cs&a-; Dr C Sahaya Kingsly, Cse-Mlmce; Dr. Chunduru Madhava, Cse-Klef; Dr. L. Bhagyalakshmi, Ece-Rec; and Dr. Sanjay Kumar Suman, Cse-Cs-Sscet on June 04, 2026, for Latency-Aware Cross-Attention Hybrid Transformer-Cnn System For Adaptive Multimodal Biometric Authentication.
Inventors include Venkatarathnam Korukonda; Kadali Satyanarayana; Gunavardini V, Cse-Bit; Vikash Sawan, Cs&a-Glau; Dr C Sahaya Kingsly, Cse-Mlmce; Dr. Chunduru Madhava, Cse-Klef; Dr. L. Bhagyalakshmi, Ece-Rec; and Dr. Sanjay Kumar Suman, Cse-Cs-Sscet.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: A novel cross-attention fusion mechanism dynamically correlates multiple biometric modalities including facial recognition, fingerprints, iris patterns, voice signals, behavioral biometrics, and palm vein characteristics to improve authentication accuracy, spoof resistance, and fault tolerance. The system further incorporates a latency- aware optimization framework that continuously monitors computational workload, network latency, memory utilization, and device capability to dynamically balance authentication speed and recognition precision. An adaptive modality prioritization engine intelligently emphasizes high-confidence biometric inputs while suppressing noisy or unreliable modalities under challenging environmental conditions such as illumination variation, sensor failure, motion blur, and background interference. The invention additionally employs anti-spoofing intelligence, feature compression algorithms, edge-cloud collaborative processing, and continuous adaptive learning techniques to enhance scalability, security, and computational efficiency.
Disclaimer: Curated by HT Syndication.
