Shopper Recognition AI
FluxVision is a property artificial intelligence (AI) technology devised by RetailFlux. Our technology is easily adapted to the existing CCTV systems and starts processing in-store video on the fly. Our AI model acts like a human brain and discriminates moving objects in the scene. That is, our AI technology can recognize individuals based their appearance and behavioral movement patterns. In addition to human objects, FluxVision AI model can swiftly recognize non-human objects such as shopping carts, baby trolleys and shadows. Thanks to FluxVision’s advanced recognition capabilities, our analytic apps indisputably outperforms industry leading 3D stereo sensors in terms of both accuracy and cost.
In addition to visual recognition capabilities, in FluxVision AI, cameras can intelligently learn the scene parameters empowering an AI based auto calibration model. In this way, FluxVision cameras adapt themselves automatically to varying environmental factors such as shadowing and lighting. Furthermore, another AI model is available for analyzing the behavior of detected people. That is, FluxVision platform can discriminate the store employees from the shoppers by monitoring people’s behavior patterns.
Beating 3D stereo sensors
Existing 3D cameras in the market are employed to filter out particular object such as carts and children by using depth information. On the other hand, FluxVision AI analytics can both filter out non-human objects and most importantly recognize unique people identities. In fact, FluxVision platform can sense identity of the subjects and track them across multiple passages. Therefore our clients enjoy extended insights provided by our solutions.
FluxVision brings extended retail insights with a lower hardware implementation costs. All kind of CCTV hardware including IP cameras, DVR and NVR devices can be employed in our platform. However, 3D stereo cameras are at least 5 times more expensive than a typical IP camera. Another major pitfall in 3D stereo cameras is that they have very low field of view capability. To be more precise, one needs to use twice as many as cameras in order to cover the same area with IP cameras.
By deploying FluxVision apps into multiple locations inside the same building, shoppers can be also uniquely tracked across distinct store departments. The tracking of people is always anonymous and EU-GDPR safe.
Superior to All
The evolution of retail analytics technologies is depicted below. See visual comparison of shopper sensing technologies.
An object is moving.
An object with height 175 cm is moving.
A human is moving. The human has visual attributes including blonde hair, white pants, red shirt, etc.
Want to know more about FluxVision and its applications?
Drop us a message!