As posted on Blippar’s machine learning tech can identify cars better than you can
Do you think you know cars?
Well, Blippar’s new machine learning technology is ready to take you on, as the augmented reality/visual search company is today announcing automotive recognition tech.
In other words, Blippar’s AI can identify the make, model and year of any U.S. car made in 2000 or after, even if the car is moving under 15mph.
Blippar originally launched as an AR platform for brands and publishers. Using a little tag (a Blipp), brands could identify content like a label of a Ketchup bottle or an ad in a magazine that users could scan with their phones to reveal extra augmented reality content.
The company has since pivoted to focus on visual search. There are plenty of things you see in the real world that are difficult to describe via text in a Google Search, such as flowers or items of clothing or an unfamiliar animal.
The company has spent the last year building out that visual search engine to identify generic objects — a table, a chair, a cup, and so on — and has laid the foundation to dive into specific verticals for visual search.
That begins with this automotive identification technology.
The technology will live within the Blippar app for folks who want to play around with it, offering information around make, model, year, average review rating, and a 360-degree view of the car, both inside and out. But the larger play comes in the form of an API, also launching today.
Secondhand sellers and insurance companies can build in this automotive identification technology into their own apps and pay on a performance basis to enhance their own businesses.
The tech has over 97.7 percent accuracy in recognizing vehicles, and Blippar says the technology is beyond what most humans can identify by sight.
We can expect to see Blippar roll out this type of technology, and accompanying API, across a number of sectors over the next year. In fact, CEO Rish Mitra mentioned that fashion is coming soon.
Blippar has raised a total of $99 million in funding from the likes of Qualcomm Ventures and Khazanah Nasional, according to Crunchbase.