MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641096879 A) filed by Vellore Institute Of Technology on August 11, 2026, for System For Mental Health State Prediction Using Digital Twin Encoding.
Inventors include Madhu Viswanatham V; and Lekha S.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: ABSTRACT SYSTEM FOR MENTAL HEALTH STATE PREDICTION USING DIGITAL TWIN ENCODING A system (100) for mental health state prediction comprises a data acquisition module (102) receiving physiological data comprising heart rate, blood pressure, sleep duration, and step count. A stress classification module (104) with a Random Forest classifier classifies the physiological data into stress levels and converts each into a signal vector. A digital twin encoder (106) with an autoencoder receives the signal vector and generates a compressed encoded memory representation of stress patterns. A temporal trajectory predictor (108) with a Long Short-Term Memory (LSTM) network receives physiological data over a plurality of days and the encoded memory representation, outputting a probability distribution across mental health states. A hybrid decision module (110) receives a prediction confidence score, compares the score to a threshold, applies a rule-based classifier when below the threshold, and outputs a mental health state prediction. A display interface (112) presents the prediction. (FIG. 1)
Disclaimer: Curated by HT Syndication.
