Computer vision and deep learning help smartphone app interpret foreign road signs

Feb. 7, 2017
Swedish startup company Mapillary—which aims to make the world accessible to everyone via crowd-sourced photos—has announced that it has expanded the coverage of its iOS and Android app to be able to recognize traffic signs from more than 60 countries using computer vision and deep learning technology.

Swedish startup company Mapillary—which aims to make the world accessible to everyone via crowd-sourced photos—has announced that it has expanded the coverage of its iOS and Android app to be able to recognize traffic signs from more than 60 countries using computer vision and deep learning technology.

Two years ago, the app was able to recognize traffic signs in a few European countries and the United States, but today, the app can recognize more than 500 signs in over 60 countries, including China, South Africa, Japan, Mexico, Slovenia, and more.

Mapillary’s app automatically finds and recognizes objects in geotagged photos submitted from users around the globe. To enable traffic sign recognition, the company uses the deep neural networks, and requires data about the appearances of traffic signs around the world, along with their corresponding meanings.

"With images contributed from all of the world by our community, we are in possession of an extremely diverse dataset to train our algorithms on," said the company in a blog post.

Because of the similarity in many of the signs used around the world, Mapillary decided to merge traffic signs into appearance groups. Each appearance group represents a set of sign appearances with little variation between different countries. This way, says the company, they are able to reduce the number of traffic sign classes for the deep neural networks to analyze.

"As a result, we have successfully trained fast and accurate deep neural networks for hundreds of traffic sign classes," said the company.

Soon developers will be able to access and use the updated signs through the company’s application programming interface. Additionally, the OpenStreetMap community can also access the data in iD editor. Moving forward, Mapillary is aiming to extend the traffic sign recognition to more countries, adding further support for complementary signs and optical character recognition (OCR) for textual signs. The company is seeking input on this update via comment,tweet, or email.

The app is available on Android (4.0 or later) and iOS (iOS 8 and later), and the following external cameras can be controlled from the iOS app: GoPro Hero3 Silver/Black, GoPro Hero4 Silver/Black, Garmin Virb X/XE. The following 360° cameras can also be controlled with the iOS app: LG360 cam (R105), Ricoh Theta S, Giroptic, Open Spherical Camera v1 and v2 compatible cameras. External camera app support for Android is forthcoming.

View the Mapillary press release.

Share your vision-related news by contacting James Carroll, Senior Web Editor, Vision Systems Design

To receive news like this in your inbox,
click here.

Join our LinkedIn group | Like us on Facebook | Follow us on Twitter

Learn more: search the Vision Systems Design Buyer's Guide for companies, new products, press releases, and videos

About the Author

James Carroll

Former VSD Editor James Carroll joined the team 2013.  Carroll covered machine vision and imaging from numerous angles, including application stories, industry news, market updates, and new products. In addition to writing and editing articles, Carroll managed the Innovators Awards program and webcasts.

Voice Your Opinion

To join the conversation, and become an exclusive member of Vision Systems Design, create an account today!