MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202611073732 A) filed by Upes on June 13, 2026, for A System For Edge-Native Retail Sentiment Analytics, Region-Of-Interest-Based Customer Behavior Correlation, Operational Alert Generation.
Inventors include Dr. Nadeem Yousuf Khanday; and Samyaak Gharat.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: The invention relates to retail video analytics and edge-based operational intelligence, and more particularly to a system and method for privacy-preserving sentiment detection, region-of-interest correlation, and real-time intervention generation. At least one optical sensor acquires live video from a monitored retail zone, and a processing module extracts a predefined region of interest, detects faces therein, generates a nonreversible privacy-preserving facial embedding without retaining raw facial imagery, and applies machine-learned micro-expression analysis to classify an emotion state with confidence. The processing module correlates the classified emotion state with the region of interest and a dwell metric, and determines whether persistent negative sentiment exceeding a dwell-duration threshold satisfies configurable intervention logic, optionally using composite operational metrics. An output component generates structured alerts for operational staff and stores or transmits only zone-linked, non-image structured metadata, enabling scalable edge-native or cloud-retrofit deployment and external system integration..
Disclaimer: Curated by HT Syndication.
