MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202611062612 A) filed by National Institute Of Technology, Delhi on May 15, 2026, for System And Method For Privacy-Preserving Biometric Recognition And Verification Using Lensless Frequency-Domain Sensing.
Inventors include Mr. Satyam Singh; Dr. Gautam Kumar; and Dr. Sachin Agrawal.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present invention discloses a system and method for privacy-preserving biometric recognition and verification using lensless frequency-domain sensing. The system comprises a Lensless Imaging Module (101) configured to capture encoded sensor measurements corresponding to a face of a subject without generating a visually interpretable face image, a Pre-processing Unit (102) configured to generate a calibrated RGB representation, a Feature Extraction Unit (103) configured to generate a multi-resolution Discrete Cosine Transform (DCT)-domain representation including an Output DCT Tensor (305), a Machine Learning Engine (104) configured to generate normalized biometric identity embeddings using a convolutional neural network backbone (402), a Transformer Encoder (404), and an Embedding Head (405), an Enrolled Identity Database (105) configured to store enrolled biometric identity embeddings, and a Decision Module (106) configured to perform biometric identification and biometric verification. The invention enables biometric inference directly from encoded lensless sensor measurements without reconstructing, transmitting, or storing a recognizable face image.
Disclaimer: Curated by HT Syndication.
