Car make and model recognition for a better world

 

     Once there was the famous phrase "The horse is here to stay, but the automobile is only a novelty – a fad." The situation changed a lot since then, and nowadays the automotive industry is often blamed for contributing to global warming and pollution. Meanwhile, we started seeing the first promising implementations of intelligent machines, an always pursued dream of humans. There are many hopes that artificial intelligence will allow for self-driving cars, to make routes optimal and trips safer. In that, car-to-car and car-to-environment networking will take a significant role. A piece that will complement that whole framework will be the independent, intelligent systems for car tracking.

 

     The founders of Neurico were friends and colleagues for many years. Around the beginning of 2016, a customer request turned their attention towards automated extraction of specific car details from images - including make, model, and year of manufacturing. They realized the impact such an application could have for reducing automotive fraud, optimizing traffic load, in the advertising industry and practically every service that depends on car data. Smart road cameras that allow for traffic statistics, intelligent billboards that show products tied to your car model, obtaining fast access to technical data of a car just by photographing it with your mobile phone, are only some of the applications. Even with the arrival of self-identifying smart cars, there will always be reasons for automated systems to have an independent way to confirm car details. Founders realized the shortcomings of existing car recognition systems, such as a limited number of supported models and the restrictions on the car view in the image. They also found that currently achieved accuracies could be significantly improved. Thus they created the startup Neurico and started developing a unique car make and model recognition technology. Building a highly-efficient AI system requires combining domain knowledge with advanced machine learning methods. While Neurico’s team realized its potential for these, the last and critical piece of the puzzle is to support all the know-how with adequate computing power. Here came their chance to be nominated for the IBM’s Global Entrepreneur Program which turned out to be the key to success for their project. A recently published blog article sheds light on how Neurico makes use of the IBM’s cloud. The current milestone for Neurico is to come to the first official release of their car make & model recognition API. They also started working on an embedded module that could allow for standalone recognition functionality in applications where access to the Internet is limited.

 

     Keep following the blog of Neurico to learn the latest news about their technology!

142