MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202641115456 A) filed by Sri Eshwar College Of Engineering on September 26, 2026, for A Deep Learning Framework For Stress And Emotional State Detection Using Wearable Sensor And Voice Data.
Inventors include G. G. Sreeja; M. Kawin; Dr. S. Dhamodharan; and R. Megala.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: Stress and emotional disturbances have become significant concerns affecting an individual's physical health, mental well-being, productivity, and quality of life. Traditional methods for assessing stress and emotional conditions primarily rely on self-reported questionnaires, clinical observations, and periodic assessments, which may be subjective and unsuitable for continuous real-time monitoring. To address these limitations, this work proposes a Deep Learning Framework for Stress and Emotional State Detection Using Wearable Sensor and Voice Data. The proposed framework integrates multimodal information obtained from wearable physiological sensors and human voice signals to provide an intelligent and non-invasive approach for detecting stress and emotional states. Wearable sensors are used to collect physiological parameters such as heart rate, heart rate variability, skin temperature, electrodermal activity, and physical activity patterns. Simultaneously, voice data are analyzed to extract relevant acoustic characteristics, including pitch, energy, speech rate, and frequency-based features. The collected data undergo preprocessing operations such as noise removal, normalization, and feature representation to improve data quality. Deep learning models are then employed to automatically learn complex physiological and vocal patterns associated with different psychological conditions. A multimodal feature fusion mechanism combines information obtained from sensor and voice data to enhance the accuracy and reliability of stress and emotion detection. The framework is designed to classify different stress levels, such as low, moderate, and high stress, and recognize emotional states including happiness, sadness, anger, fear, relaxation, and neutrality. The proposed system supports continuous and potentially real-time monitoring, enabling early identification of abnormal stress and emotional patterns. The framework can be applied in healthcare monitoring, workplace wellness, student well-being assessment, elderly care, and personalized digital health systems.
Disclaimer: Curated by HT Syndication.
