This article describes the conference sessionHoloLens as a tool for computer vision research, held September 8, 2018 at theEuropean Conference on Computer Vision (ECCV) 2018.

Microsoft HoloLens is the world’s first self-contained, holographic computer, but it’s also a potent computer vision research device. Application code can access audio and video streams and surface meshes, all in a world coordinate space maintained by HoloLens’ highly accurate head-tracking. This tutorial session featured the new HoloLensResearch Mode capability.

The tutorial showed how to access the raw head-tracking and depth sensor data streams, and use the intrinsic and extrinsic parameters of each stream. The session also demonstrated recent advances in time-of-flight depth-sensing technologies in the Kinect for Azure project.

Session attendees got a good sense of how to use HoloLens for a range of computer vision research tasks. Attendees received materials to help them quickly get started using HoloLens.

The following image shows a sample HoloLens application that displays any of the six Research Mode streams in real time.

Session organizers

• Marc Pollefeys
• Shivkumar Swaminathan
• Johannes Schoenberger
• Andrew Fitzgibbon

Session schedule

• 0900 – Introduction
• 0930 – HoloLens Research Mode
• 1030 – Coffee break
• 1100 – Applications & Demos
• 1130 – Kinect for Azure Depth Sensor
• 1200 – Q & A
• 1230 – End

For more information and conference proceedings, seeECCV 2018.