AI Kiosk

Deep Runner provides an artificial intelligence solution optimized for AI kiosk configuration. This device can automatically recognize the age, gender, facial expression, etc. of the customer in front of the kiosk, and provides additional customer optimization services on the verified platform. Various kiosk application services can be developed that provide services while interacting closely with customers at the minimum cost and in the shortest time.

<AI Kiosk Demo Video>

Applications

• smart signage system
• smart billboard
• ticket sales booth
• smart vending machine
• smart access control

Configurations

Input/output

• Camera input

            MIPI CSI-2 

            USB Camera (UVC) 

            IP camera 

• Output

            Self-save of recognition results 

            Real-time server communication via TCP/IP 

            Real-time server communication via TCP/IP 

Features

• Using the best AI technology, the age, gender, and expression of a moving person's face are recognized at the same time and the recognition result is transmitted to the service module.

• Developers can further train the necessary deep learning models to enhance service scenarios, and for this, we provide the necessary development environment and technical support. For example, you can add a customer's gesture to a conversation scenario and implement a specified action when the customer makes the special gesture.

• As a built-in AI algorithm, the MobileNet 1.4x classification model with the highest recognition rate and the SSD object recognition model are supported. Floating point computations are performed instead of integer, so that the recognition rate is not degraded during the computation process.

• Built-in signage set-top communication function makes it easy to link with the signage set-top in the kiosk.

• By default, the server web protocol is installed so that it can be linked with the service IT system without additional development, and it provides technical support so that customized interworking functions can be easily added if necessary.

• If necessary, the service scenario can be implemented as a C program on the Linux operating system in the device. Sample programs and technical support are available.

• It provides an environment optimized for mini USB cameras.

• As a small board of 85 x 56 mm, a slim stand can be realized, and it can be easily installed in any environment.