Applications

Measure activity on the shop floor.

Retail is YUAN's newest application area. The proposed system applies people analytics and edge inference already used in city and campus deployments to video from a store's existing cameras.

Ceiling security camera overlooking a retail aisle

The opportunity

Measure in-store traffic with video analytics.

Online retail can measure visits and navigation paths. Physical stores can use existing camera feeds for on-site counting and movement analysis.

YUAN proposes compact NVIDIA Jetson-powered edge platforms that analyze store video locally and output configured counts and events.

People analytics

Measure footfall, dwell time, and movement.

SmartFlow provides people counting and movement analytics. In a retail configuration, it can output entry counts, movement paths, dwell estimates, and heat maps.

An edge deployment can retain raw video on site while forwarding aggregate counts and configured events.

Overhead view of shoppers moving through a store aisle

Loss prevention

Run loss-prevention analysis on site.

YUAN uses anomaly-detection workflows in public-safety systems. A retail implementation could be configured to detect unusual movement patterns and after-hours activity.

Staff can review configured event clips instead of manually scanning all recorded footage.

Security workstation displaying multiple store camera feeds

The honest note

No published retail case study yet.

YUAN does not yet publish retail case studies. Candidate components include Pandora edge-AI kits from 67 to 157 TOPS, VPP series platforms, the EYE AI camera, and the NexVDO SDK. These products are used for related video-analysis work in adjacent industries.

Retail teams would need to validate camera placement, privacy requirements, model performance, and alert workflows in their own environment before deployment.