Retail video analytics solutions use artificial intelligence (AI) technology to understand video streams/feeds, offering brick-and-mortar retailers the potential for a similar level of shopper insights as those used by the most sophisticated e-commerce sites.
The more retailers can learn about their shoppers, the better they can serve them – leading to improved sales and profits.
The Retail Video Analytics Platform from SMART Embedded Computing combines real-time analytics, interactive customized promotions and employee nudges to increase sales and improve operational efficiency.
You can integrate with existing security cameras to cost-effectively enhance your customer experience and streamline operational costs.
The system encrypts all data and does not collect personally-identifiable information.
By measuring visual data and comparing trends over time, retailers can enhance store performance and operations:
SMART Embedded Computing is working with leading video analytics AI software providers, to create a set of scalable solutions to transform the way retailers use, store and access their visual data.
The SMART Edge Server SE1700, with integrated GPU AI inference engine and hosted video analytics software, can process up to 8 cameras using state-of-the-art data analysis software providing the ability to:
The SMART EC Video Analytics Platform for retail applications provides a real-time dashboard with immediate access to information such as:
The solution is designed to capture what is considered non-personally identifiable information. It collects pixel-level environmental information on selected objects, such as a car driving into a gas station or a person walking into a building.
Our AI platform aggregates and analyzes that environmental information against environmental information collected in previous encounters of similar class objects to recognize patterns. It then generates confidence interval data as to whether an object is recognized as the same class of object as in one or more previous encounters.
The confidence interval data cannot be used to distinguish, identify or trace an individual’s identity (no biometric records, etc.), and is only used to recognize, track, and help understand behaviors of classes objects.