In the year that businessman Steve Jobs died and the work of philosopher Walter Benjamin entered the public domain, the frantic eyes of consumers could only be dazzled by one thing: the recently launched Nintendo 3DS. The 2011 video game was the newest addition in a long line of handheld consoles from Nintendo, the last addition of which came in 2004 with the Nintendo DS.
Apart from some improvements in processing power and a new slider button, the console was very similar to its predecessor. The bet, however, was on something else: the new autosterioscopic 3D screen – which, without the need for special glasses, simulated a depth effect.
Technology at the time was still limited. The effect of autosterioscopic 3D screens only works in a very restricted space – in the case of video games, a distance approximately between 25 centimeters and 40 centimeters from the screen – and any wrong movement or angle would break the illusion.
Now, taking augmented reality technologies a step further, Chinese researchers have developed a new display which uses artificial intelligence to track gaze and shoot high-definition images directly into viewers’ eyes. THE EyeReal promises to create images that remain deep and sharp even with rapid head movements.
This all happens because it gets around a fundamental problem of current displays: oblivious to the position of the observer, they shoot the image everywhere. As far as no one can see.
How 3D screens work
To create a three-dimensional image of the world, our eyes see stereoscopically. That is, each of our eyeballs, minimally distant from each other, captures a slightly different angle of the world, which our brain puts together to create an image with depth. The name of this effect is parallax.
Autosterioscopic screens simulate this by showing a slightly different image to each eye, the visibility of which depends on the angles at which they are viewed. To make this possible, there are some methods.
Share this article via:
The 3DS, for example, uses a “parallax barrier”, that is, very thin opaque plates that separate, in the spaces between one and the other, what each eye sees. For the effect to work, however, the observer needs to have their nose exactly at the point of intersection between these two image zones.
In short, autosterioscopic screens do not know exactly where to display each angle of the image, and, in the absence of a specific direction, they need to display both images across the viewer’s entire field of vision.
In turn, looking directly at the eye of the beholder, the EyeReal needs to make a very delicate calculation in which the displays pasts generally failed. The slightest discrepancy in the image received by each eye is enough to cause discomfort, eye fatigue and headaches, in addition to breaking the viewer’s immersion.
The new technology solves this accuracy problem by using a deep-learningthat is, a computational learning system that uses neural networks to, in this case, learn the best way to track our gaze. The AI finds the eye, calculates its position in real time and the display shoots the image in a small region around each orbit, creating a view that, even when the observer moves, remains deep, sharp and immersive.
All this using components accessible to most consumers. However, the system is still a prototype and has not been made available for purchase. Even so, it remains to be seen whether the fascination that occurred in 2011 with the launch of the 3DS can reappear now, in a new, even more immersive dimension.
Scientists describe the innovation in a paper recently published in the journal Nature.
