Cubic Motion announced a new service supplying advanced facial tracking to clients and partner studios who require rapid and efficient analysis of facial video capture or other data. Powered by new breakthroughs in computer vision technology, the service is immediately available for a wide variety of capture methods, including single and multi-camera head-mounted systems, free-standing cameras, and 4d data. Tracking is completely configurable; required features can be specified by the client to get precisely the data needed for their own systems. Features available include: fiduciary points, contours, texture analysis and even automated translation to FACS-like measurements. Cubic Motion can also supply accurately estimated predictions of partially hidden data such as upper and lower teeth. Maximum tracking precision is determined only by the fidelity of the source video or other data.
Dr. Gareth Edwards, CEO of Cubic Motion explained “There are many choices when it comes to facial capture. In fact, Cubic’s services are sometimes misunderstood - we’re not another capture option, rather, we’re here to make sure clients get the very best results out of any given capture and - if asked - to help our clients to choose the right type of capture. Of course, for developers, ‘best’ includes scalability and cost. One thing we know for sure is that game developers already have too much to do. They don’t want to waste precious resource tracking and analyzing video or other data. Not only is it expensive, it’s time-consuming and very hard to maintain quality across a large team.
We’re working on a large slate of next-gen titles, and in most cases, our clients also rely on our expertise to transfer performance (as fully editable animator-friendly data) to their rigs, delivering the most efficient high-end animation in the industry. However, we also recognize that some teams wish to build their own solvers, powered by great data. Either way, two decisions are vital: first - choose the capture source that has the right complexity/cost balance for the rigs you’re driving. Second, be sure to get the best possible measurements you can from that data – it’s crazy to spend a fortune on the most sophisticated capture you can buy and then reduce the data to a few dots before maybe stuffing it through some rough animation tool or process – why bother at all?
We can help clients to choose the most cost-effective capture for any given rig, by demonstrating exactly what measurements can be made from the data, and indeed how to solve it to your rig. In many pipelines, a single-camera head-mounted system works just fine (Mocap Design’s wireless units are a great choice), but you can send data of arbitrary resolution and number of channels, and we’ll make whatever feasible set of measurements your system requires. This is all made possible by some amazing breakthroughs in computer vision - of which you’ll see a lot more in the coming six months – including live tracking and live animation at a fidelity we’ve never seen before.”
Source: Cubic Motion