Press "Enter" to skip to content

Nvidia is developing gaze tracking system using LED sensors

Top virtual reality headsets makers all over the world agree that gaze tracking technique is going to be most important and fundamental technique to the next generation of Virtual Reality hardware. Gaze tracking system has the calibre to sense and detect the position of eye in real time. Thus the technology allows a computer to offer cursor control without the need for head and movements and optimize detail rendering. However, currently the gaze tracking hardware system is not small. And the system is expensive too. Therefore to enhance this disadvantages of gaze tracking hardware, experts at Nvidia have come up with a new idea and solution. That idea could aid the technology to help experts as well as become more widespread. Thus the new gaze tracking system by Nvidia uses some common features like LEDs. LEDs have some important feature like ability to emit as well sense light. Therefore this ability of new gaze tracking system will help to simplify the process of detecting the position of eye relatively to a display screen. In addition, new gaze tracking device by Nvidia uses a ring of infrared LEDs like other gaze tracing devices. This ring enable to project light into the eye which is unseen. But LEDs are also has the calibre of selection of color sensing from the same location. This system is the smallest gaze tracking system developed yet according to the researchers. This system is cost effective as well. Also sampling rates and accuracy are matching with today’s most common solutions. In one system of gaze tracking, Nvidia is using total number of nine LEDs per eye. This prototype includes six sensing the light and three emitting IR light. Whereas the second prototype of the system include six LEDs per eye. This LEDs are used as light sensors as well as light sources. LEDs are used in the device because it consumes less power. It also rely on simple software and simple controller hardware compare to others. In addition, they overall latency. Furthermore, LEDs enable to lower the number cameras needed for the headset. It further remove an extra image processing block need within the pipeline of headset. Researcher are not so sure about the device if is suitable for psychological research, neurological applications and reading in spite of its solutions is performant enough to work for typical VR applications. But the LED device has good median angular error of 07 degrees. Mean angular error is as low as 11 degrees. In addition, other alternatives based on the camera can deliver very high accuracy results with error levels under 0.5 degrees. The initial calibration phase of the system made by Nvidia noted that it is comparatively longer than other solutions. In many cases “look here, here, here, and here” techniques uses in a system in order to match up with eye as well as must recalibrate if the wearer’s face moves relative to the sensing hardware. The gaze tracking system by Nvidia is still in the stage of prototype and it is not prepared fully to challenge the Tobii solutions found in Pico VR headsets and HTC.