Lidar sensors manufacturer Ouster has partnered with data labelling companies Playment and Scale.AI to support the implementation of deep learning models in self-driving vehicles.

The partnership aims to streamline data annotation and facilitate the development of autonomous vehicles.

It will leverage Ouster’s multi-beam flash lidar technology that provides intensity, range and ambient images where all 2D pixel corresponds to a 3D point.

The agreement will also enable the data labelling partners to translate between 2D and 3D point clouds supporting the machine learning algorithms in self-driving cars.

A joint team has already been working to devise a common data format and labelling process, utilising the image-like output from Ouster’s lidar sensors and the capabilities of Playment and Scale.AI’s labelling toolchains.

“We’re revolutionising the hardware, software, and services that turn 3D data into actionable inference, with the goal of democratising access to lidar technology.”

The approach is said to eliminate key bottlenecks and time-consuming aspects of the data labelling process.

Ouster CEO Angus Pacala said: “Both Playment and Scale.AI are on the cutting edge of data labelling technology and they’ve re-imagined their lidar toolchains around our senor’s unique 2D-3D technology to improve accuracy, reduce costs, and eliminate friction.

“Together with our partners, we’re revolutionising the hardware, software, and services that turn 3D data into actionable inference, with the goal of democratising access to lidar technology.”

Initial results of the partnership indicate significant simplification in the process required to feed lidar data through the data annotation APIs.

It also promises to reduce data labelling costs by up to 50%.

Playment cofounder Ajinkya Malasane said: “We think that this partnership will accelerate the timeline toward safer, more reliable, and more fully autonomous vehicles on the road.

“We can do so much more when we align our efforts with lidar leaders like Ouster.”