MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202611047946 A) filed by Ravindra Kumar Sharma on April 15, 2026, for Conversational Awareness Engine.
Inventor includes Ravindra Kumar Sharma.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: A multimodal nonverbal social state detection system and method for real-time interpretation of the full spectrum of human nonverbal communication during conversational and social interaction contexts. The system captures and fuses nine distinct nonverbal signal channels: (1) wristband physiological signals including photoplethysmography (PPG), electrodermal activity (EDA), skin temperature, and accelerometry; (2) facial expression analysis via full Facial Action Coding System (FACS) Action Unit extraction across 46 action units; (3) micro-expression detection using a highframe-rate imaging unit operating at not less than 120 frames per second within temporal windows of 40 to 200 milliseconds; (4) eyecommunication analysis comprising gaze direction, pupil dilation, blink rate, saccadic movement, and sustained eye-contact duration; (5) head movement kinematics including nodding, head-shaking, headtilt, and head-retraction events; (6) whole-body posture estimation via skeletal keypoint analysis from a dual-camera configuration: (7) involuntary kinematic micro-signal analysis using a jerk tensor J(t) = d'P{t)/dt³ computed for body keypoints detected across both imaging units and from wristband accelerometry, specifically capturing the sub-300ms involuntary bodily micro- movements that communicate appreciation, disgust, scorn, engagement, and affective states below the threshold of conscious control; (8) communicative gesture recognition across hand emblems, illustrators, adaptors, and regulators; and (9) a composite social vibe vector S(t) that fuses all foregoing modalities into a structured representation of the social-emotional state of the subject. The dual- camera architecture - comprising a primary frontal imaging unit facing the subject and a secondary head-worn imaging unit providing a first-personperspective view of the subject's body - is a novel configuration that enables simultaneous full-face and full-body coverage without requiring the subject to remain stationary. An on-device privacypreserving processing architecture ensures no raw video, raw biosignal, or raw pixel data is transmitted off-device; only abstract feature vectors are processed in Stage 2. The social vibe vector is output via an API for consumption by conversational AI systems, therapeutic tools, and educational applications.
Disclaimer: Curated by HT Syndication.
