Imagine strapping on a VR headset for the first time and getting immersed in a virtual world, and then feeling nauseous. Not everyone experiences the phenomenon, but for those who do suffer from motion sickness in VR, it can be a crippling experience that ruins what could otherwise be an awesome adventure. VR developers have spent considerable time and research to reduce the effect, and interestingly enough one of the solutions has to do with artificial intelligence.
AI has been around for decades, though only in recent years as it penetrated so many different market segments, everything from self-driving cars and social media schemes (like recognizing faces in uploaded photos), to smart phones and smart speakers. As it pertains to VR, LG Display and a team from Sogang University have developed what they claim is the world’s first AI-based content creation technology.
Here’s the thing about VR—dizziness and motion sickness are largely the result of photon latency and motion blur. Unfortunately, the effects are exacerbated at higher resolutions, and of course higher resolutions are desirable for more realistic settings and to combat the screen door effect. Higher resolutions require more processing power, and while there are obvious benefits to increasing the number of pixels, it comes at the expense of photon latency and motion blur.
That’s where this new AI technology comes into play. Using an algorithm, it can take low resolution images and output ultra-high resolution ones, and do it in real time. The technology also leverages deep learning to make the conversion possible without having to depend on external memory devices.
“This study by LG Display and Sogang University is quite meaningful in that this study developed a semiconductor which accelerates with low power realized through AI without an expensive GPU in a VR device,” said Kang Seok-ju, a professor of the Department of Electronics Engineering who’s been carrying out the study for the past three years.
By turning lower resolution images into higher resolution ones, this AI scheme can significantly reduce the delay between when a VR user moves his or head head and the motion that appears inside the headset. This in turn can have a big impact on lessening motion sickness.
It will be interesting to see if VR makers ultimately adopt this sort of thing, or focus their efforts more on increasingly powerful hardware instead.