Published On: Wed, Aug 2nd, 2017

NVIDIA and Remedy Partner Up To Streamline Production-Level Facial Performance Capture Using Deep Convolutional Neural Networks

Remedy Entertainment has partnered adult with NVIDIA to streamline production-level facial opening constraint regulating low convolutional neural networks.

As reported by Ars Technica, NVIDIA presented a “deep training framework” with Quantum Break developer Remedy Entertainment during this year’s Siggraph event. A video was common as well, that can be watched below:

df7upervoaaabv4Related AMD’s Biggest GPU “Radeon RX Vega” Launch Event In Two Years Won’t Be Livestreamed – Capsaicin Siggraph Exclusive To Press and Analysts Only

We benefaction a real-time low training horizon for video-based facial opening capture—the unenlightened 3D tracking of an actor’s face given a monocular video. Our tube starts with accurately capturing a theme regulating a high-end prolongation facial constraint tube formed on multi-view stereo tracking and artist-enhanced animations. With 5–10 mins of prisoner footage, we sight a convolutional neural network to furnish high-quality output, including self-occluded regions, from a monocular video method of that subject. Since this 3D facial opening constraint is entirely automated, a complement can drastically revoke a volume of labor concerned in a growth of complicated narrative-driven video games or films involving picturesque digital doubles of actors and potentially hours of charcterised discourse per character. We review a formula with several state-of-the-art monocular real-time facial constraint techniques and denote constrained animation deduction in severe areas such as eyes and lips.

While a technique appears to have limitations, a presented is means to grasp “high levels of correctness for facial opening constraint while drastically shortening a volume of primer work concerned in a prolongation setting”. NVIDIA and Remedy write that a technique has “proven accurate adequate to be practical in an arriving diversion prolongation formed on consummate pre-production contrast while also requiring most reduction labor than other stream facial opening constraint pipelines used in diversion production.”

nvidia-tesla-v100-volta-gpu_2Related NVIDIA’s CEO Gives Away Their Mighty Volta GPU Based Tesla V100 AI Accelerators To Top 15 AI Research Institutions

More about this technique can be found in a paper that has been submitted by Remedy and NVIDIA.

About the Author

Leave a comment

XHTML: You can use these html tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>