Abstract

Augmented reality (AR) glasses and other head-mounted wearable devices (such as AI glasses or VR/XR headsets) have competing hard-to-meet requirements with respect to battery life, weight, cost, aesthetics, etc. AR glasses include world facing cameras for perception tasks such as head tracking (HeT), hand tracking (HaT), 3D reconstruction (3DR), etc., as well as RGB cameras for high-resolution image/video capture and artificial intelligence tasks. However, having multiple cameras on AR glasses adds to the weight, imposes greater power requirements, impacts device aesthetics, and requires more computational capability to process more images. This disclosure describes augmented reality glasses with a lower number of cameras that deliver functionality without sacrificing the user experience. Per the techniques, the AR glasses include one monochromic camera and one RGB camera. The RGB camera capture is concurrently used for perception tasks (HeT/HaT/3DR) and other tasks traditionally performed with the monochromic camera. In particular, the capture mode of the RGB camera is configured to dynamically alternate between (or otherwise interspersed) high resolution image capture (e.g. 12MP) for visual tasks such as AI assistance or video recording and low resolution capture (e.g. VGA) for perception tasks. In another implementation, the RGB camera is configured to output both high resolution frames and low resolution frames concurrently in multi-stream mode to be used for perception tasks and AI/videography use cases.

Creative Commons License

Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.

Share

COinS