Imagine spilling a Boxes full of Lego bricks on the table. Now — take a leap with me — don’t imagine yours augmented reality glasses. The camera in the AR glasses will immediately start cataloging all the different bricks in front of you, ranging from different shapes to colors, making suggestions for models you can build with the pieces you have. But wait, there’s someone at the door. You go check it out and come back. Thankfully, your glasses don’t need to rescan all those pieces. AR know they are sitting on the table that you have left them.
The ability to continuously remember scanned real-life objects is the heart of the new AR software platform called Perceptus from Singulos Studies. Perceptus can keep those objects in memory even when the camera is no longer directly looking at the scene. As you step forward to answer the door, the Perceptus platform continues to think about what else you can build with the pieces on the table. It doesn’t stop working just because you are no longer looking at the fragments.
“When we’re in an AR space, we’re not looking at the entire room at once, but only a part of it,” said Brad Quinton, CEO of Singulos Research. “As humans, we have no trouble thinking that there are things that exist that we cannot see at the moment because we have seen them before and we miss them. Once you have an AR that can understand what’s around you, the AR can be active and proactively do things for you. “
At least, that’s the idea. Perceptus acts as a layer on top of existing AR technologies like Apple’s ARKit or by Google ARCore, which developers use today to create AR apps. But a lot needs to happen behind the scenes before this can work on your smartphone or tablet.
App developers provide Singulos Research with 3D models of Lego bricks — or any other object — they want Perceptus to detect. The platform then uses a kind of machine learning process, where it studies all the different ways it might expect to see the object in the real world, with different lighting conditions, on different surfaces etc. Perceptus is then layered on top of the developer’s application, allowing it to use this new object comprehension. It’s the developer’s job to make sure that the app actually gives you things to do with objects, the same way our imaginary Lego app might suggest things you can build with using the bricks it identifies.
The process of scanning and recognizing objects is still a lot of manual work. To get started, application developers that license the Perceptus platform will need to provide computer-aided design models of the objects they want it to remember. But those CAD models will be added to Singulos’ library, and future developers will be able to search through digital stacks to quickly find the objects they need. Before long, Quinton hopes Perceptus can identify a range of popular items — especially since “a large number of very precise 3D models” are available from video game makers. .
