MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202611092907 A) filed by Chandigarh University on July 31, 2026, for Adaptive Bayesian Audiometry Using Probabilistic Contralateral Masking And Gaussian Process Modeling.
Inventors include Harsh Preet Singh; Upasana Tiwari; Rza; Rajveer Sharma; and Niharika Sharma.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The present invention relates to an adaptive Bayesian audiometry system and method for automated pure-tone hearing assessment using probabilistic contralateral masking and uncertainty-guided decision-making. The system jointly models test-ear threshold, non-test-ear threshold, and interaural attenuation as latent stochastic functions through a multi-output Gaussian Process. An acquisition function selects the most informative tone based on posterior uncertainty during each testing cycle. The system estimates the probability of cross hearing and activates contralateral masking only when the calculated probability exceeds a predefined threshold, while dynamically optimizing masking intensity. Patient responses continuously update the posterior model to improve threshold estimation accuracy. The system further incorporates uncertainty-driven reliability verification by selectively introducing validation stimuli when posterior variance exceeds a specified limit. A posterior- constrained reporting module generates clinically consistent hearing assessment reports supported by probabilistic evidence. The invention reduces unnecessary masking, shortens examination time, improves diagnostic accuracy, enhances patient comfort, and enables reliable automated audiometric reporting.
Disclaimer: Curated by HT Syndication.
