A new artificial intelligence program called Volume is able to analyze 2D images and videos, and put objects from them into 3D spaces using augmented reality. The team behind Volume, Shirin Anlen and/or Fleisher, made a video to demonstrate the technology, as well as a few small demos on the project’s website. The demo projects on the site move with a user’s mouse, while the video depicts one of the many possible use cases in the form of a semi-interactive scene from Quentin Tarantino’s irreverent genre-defier, Pulp Fiction. In the scene, two characters dance in the videographer’s living room, and the person taking the video actually walks right next to them. The characters scale in real time in response to the camera’s movements, and their size is calculated as fairly close to what they would be in real life.
Volume uses a convolutional neural network to do its magic, which means that it could potentially be ported to a number of different AR architectures and tool sets, such as Google’s ARCore. The video demo involving characters from Pulp Fiction, for instance, is shown in real time on an Apple iPad, using Apple’s own ARKit framework as a backend. The software sets itself apart from similar solutions in the market by its endgame premise; rather than being a suite of tools meant to allow creators to craft immersive experience for end users to enjoy, it’s actually the end users who will shine with this project. Specifically, it’s going to have an end-to-end web wrapper that’s device-agnostic, allowing just about anybody to see their favorite 2D content rendered in augmented reality.
Volume is by far not the first tool of its kind, but it will be the very first one to integrate a suite of automated tools that make things as easy as possible for end-users. In a way, it’s not unlike Google’s new VR180 initiative; it takes high-tech creative solutions enabled by modern technology, and simplifies it as much as possible so that everybody can benefit from it, rather than being limited to enthusiasts willing to spend a lot of money or the highly computer-literate crowd.