Deep learning by public road driving Data set creation service for research and development Start of receiving orders

For deep learning research and development by public road driving
Start dateset creation service orders

- One-stop implementation from shooting image data on public roads to labeling for deep learning learning -

ZMP Corporation (Bunkyo-ku, Tokyo; CEO: Taniguchi Hisashi) announced today that it will be available as a new menu for service of public road data acquisition support service RoboTest (Robotest) as a learning data set for deep learning We begin to receive orders for RoboTest annotation service that performs labeling on acquired data in pixel units.

As the revitalization of the ADAS market and the development of Autonomous Driving technology are progressing, as a method of safe driving support, utilizing deep learning, recognition precision such as a travelable area, a preceding vehicle, a surrounding vehicle, a pedestrian, a signal, a sign, a white line. Attempts to raise awareness have been actively studied. In image deep learning, it is necessary for tens of thousands sheets of data accurately labeled objects and road areas in the image as learning data.

From 2012, we will provide public road data acquisition support service "RoboTest" which acquires image data of the outside environment when driving by public roads together with vehicle information (vehicle speed and pedal depression state) etc. RoboTest has offered it as a one-stop service that delivers data according to customer's request after designing and constructing a data logging system from upstream processes such as planning acquisition of driving data, arranging the subject driver and arranging data and managing it It was.

The RoboTest annotation service to be launched this time will implement the data acquired by RoboTest as a learning data set all the way to various labeling so that customers can agree to the necessary data specifications at the beginning so that the system and driver.You will be able to obtain learning data without worrying about management etc. In addition to labeling from the public road data collection service, it is also possible to label the image data acquired by you.

【Related product information】

Data set making service for deep learning R & D for public road driving

【Exhibition Information】

Automotive Technology Exhibition Engine and Technology Exhibition for People and the Car 2017 Yokohama
Period: Wednesday, May 24 - Friday, June 26, 2017 Venue: Pacifico Yokohama - Exhibition Hall
Booth number: 102
Press correspondence: Wednesday, May 24 (Wednesday) 9:30 to 9:50 at the ZMP booth
Exhibition homepage:

[Press Release Article]




Life Robot Brothers

ZMPが描く自動運転で実現する未来 ~ZMP Vision~