MUMBAI, India, Feb. 27 -- Intellectual Property India has published a patent application (202641019022 A) filed by Nitte, Mangaluru, Karnataka, on Feb. 19, for 'an intelligent calibrated meta-learning and meta-reinforcement learning system for micro-expression recognition.'
Inventor(s) include Chandra Singh; Dr. Shwetha N; Dr Rangaswamy Y; Dr. Gangadhar N; Dr. Virupaxi B Dalal; Dr. Aravind Jadhav; and Sangeetha N.
The application for the patent was published on Feb. 27, under issue no. 09/2026.
According to the abstract released by the Intellectual Property India: "Disclosed herein is an intelligent calibrated meta-learning and meta-reinforcement learning system for micro-expression recognition (100) comprises an input acquisition module (102) configured to receive high-frame rate facial video sequences. The system also includes a feature extraction module (104) configured to extract spatial facial features from each frame. The system also includes a calibrated regression and image alignment module (106) configured to align micro-expression feature. The system also includes a meta-learning module (108) implementing a model-agnostic meta-learning framework to enable rapid adaptation to new micro-expression recognition tasks using limited training data while improving cross-dataset and cross-subject generalization. The system also includes a meta-reinforcement learning module (110) modeling task adaptation as a Markov Decision Process and applying policy optimization to refine meta-parameters for stable convergence and global optimization. The system also includes an output module (112) configured to generate a recognized micro-expression category with reduced uncertainty and high accuracy."
Disclaimer: Curated by HT Syndication.
