It took about a decade, but most of us have given up on carrying around a traditional camera. Instead, we just use our phone to take pictures. But who would think we might be saying the same thing about 3-D scanning in just the next few years?
This app in development by researchers at ETH Zurich makes quick-and-dirty 3-D scanning a cinch on Android devices. You simply open the app, aim the phone at a subject, and move its perspective around following onscreen prompts. Meanwhile, the software takes several photos automatically, extrapolating the object’s geometry, and mapping bits and pieces from all of those 2-D photos onto a 3-D figure.
Building 3-D objects from several photos isn’t a new idea. Recently, we featured an artist who’d attempted to build her self-portrait using such technologies, along with another artist who’d been mapping all of the objects in his life using a Kinect.
What makes ETH Zurich’s technology different, however, is that you can build a 3-D object at its point of acquisition–you can actually capture and render an object in 3-D, right where you see it in the real world without any special equipment. (In this sense, it really is about the closest thing you can get to a 3-D Instagram).
All processing is happening in real time on your phone–displayed right on the screen in front of you–so there’s constant, immediate feedback as to different angles you need to create the map. And, at least according to the demo above, the results are just very, very good. Researchers are using a combination of tactics to drive the high accuracy: Recognizing an object’s geometry visually, as well as understanding how the camera’s relative position is changing to the object via the phone’s accelerometer/gyroscope data.
Of course, all of that’s splitting geek hairs. What really matters here is, soon, everyone will be able to walk around the world and snag any texture or shape they see. Technology like this could allow us to crowdsource every object in the world like some kind of 3-D Google Streetview. Or, you know, it could just bring a whole new fidelity to the world of selfies.
The team hopes to bring the product to market, in some form, in the 6-12 months.