Cloudian-Powered Vehicle Recognition Advertising Model in Production

Cloudian, Inc., provider of hybrid cloud object storage systems, announced that its HyperStore technology-enabled roadside digital advertising project in Tokyo has transitioned from a proof-of-concept to production use.

Spearheaded by Cloudian, Dentsu, Inc., and Intel Corporation, the project employs AI-based vehicle recognition using HyperStore’s object storage and deep learning capabilities to present relevant display ads to drivers based on vehicle make and model. The initiative is currently implemented in one smart billboard located in Roppongi, Tokyo, Japan, and is projected to expand to other areas within the city in 2017.

“This technology demonstrates how Cloudian object storage revolutionises the economics of data-intensive applications such as machine learning, making futuristic solutions now a part of our daily lives,” said Mike Tso, CEO, Cloudian. “The world’s data is growing at an exponential rate, a trend that will only accelerate as IoT and machine learning technologies gain momentum. Cloudian strives to leverage the world’s data to create powerful new experiences for businesses and consumers.“

The HyperStore-powered system utilises automatic feature extraction of traffic patterns and volume, and automatic vehicle recognition to enable the targeted advertising via roadside, digital signage. The system currently recognises over 300 unique car makes and models. Once the system has recognised a target vehicle, the smart billboard then selects an appropriate stored image to present a customised digital advertisement to the driver in half a second.

Ads are chosen for specific vehicles based on requirements set by advertisement sponsors. For example, if a beverage company is a sponsor, the programmed system will select an image of one of its products from the database when it detects an appropriate vehicle to target-such as a canned coffee or energy drink for truck drivers traveling long distances. The system does not store or archive any of the vehicle photos.

With other possible applications including traffic volume surveys through vehicle differentiation and existing traffic cameras, as well as road monitoring projects for safety and security with licence plate recognition, HyperStore provides many possible use cases for companies seeking to utilise big storage and deep learning technology for similar machine learning initiatives.

Video