fxpodcast #311: Prof. Hao Li, the new head of USC ICT Graphics

This week we speak to USC Viterbi Professor Hao Li who is going to spearhead the Vision and Graphics Lab at the USC Institute for Creative Technologies (USC ICT). This interview is somewhat of a primer for more  detailed stories on facial research here at fxguide, starting next week.

li_2014Hao Li also presented a course at SIGGRAPH 2016 on Digitizing the Human Body. Li is an expert in 3D Human Digitization and especially complex real time facial capture and animation.

Li, has been an assistant professor of Computer Science at the USC Viterbi School of Engineering and holder of a Viterbi Early Career Chair, before taking the helm at the USC ICT Graphics Lab.


ICT ‘s Graphics Lab is famous for the work that Dr. Paul Debevec and his team did in high dynamic range imaging, image-based lighting, and developed the Light Stage. Fxguide spoke recently to Dr. Debevec about his move to Google, and this week we talk to his successor. Li's agenda for ICT graphics will naturally be different, and follow his team's interest in the democratizing facial capture and animation. When Paul Debevec joined Google, he was joined by several of his fellow ICT team, so Li is not just taking over the Graphics unit but also setting a new agenda along with new staff.

With this in mind, we discuss the current state of facial capture and his excellent SIGGRAPH 2016 Course .which was co-presented by speakers from MIT.

siggraph
The SIGGRAPH 2016 course

Before joining USC in 2013, Li, a renowned expert in the field of 3D digitization and performance capture, was a lead researcher at Industrial Light & Magic/Lucasfilm, where he developed the next generation real-time facial performance capture technologies for virtual production and visual effects.

ilm

He is ranked as the top researcher on the Top 10 Leaderboard of Computer Graphics research by Microsoft Academic for the past five years and one of the 35 innovators under 35 by MIT Technology Review in 2013. Dr. Li plans to blend the expertise of ICT’s Graphics Lab, which uses high-end capture devices with his existing research from USC Viterbi, which focuses on more deployable tech capture solutions such as building a 3D avatar from a single images, or from conventional video cameras.

Li is an incredibly engaging person, bright, quick and very intelligent, - we were lucky enough to visit his new company Pinscreen, which is mentioned in the podcast but as yet is not really discussing their work publicly. Fxguide will report on it's revolutionary work as soon it gets out of what Li calls "stealth mode".

Li’s goals include expanding USC ICT’s leadership in human digitization and exploring new AR/VR platforms. We did not discuss his VR work extensively in this podcast but fxguide had covered this in the past. In addition to defense and health applications, his work aims to expand the role of augmented reality and virtual reality for new, immersive 3D teleconferencing technologies, as well as deployable performance capture technologies in sports.

As you can hear in the podcast, - Li expects to broaden the direction of research to grow the computer vision capabilities for the Graphics Lab and build new machine-learning research that can infer complex digital 3D content automatically from minimal input.

demo
Earlier research from USC, the newest paper from 2016 will be covering in an upcoming fxguide feature story

“We want to focus on real-time and data-driven algorithms that can perceive the 3D world like humans do and even make sense of it, given limited visual input. We will investigate learning-based and geometry-processing techniques that can handle large amounts of 3D data to achieve this goal", said Li.

“The most important thing,” said Li, “is to modernize the way we capture, process and analyze data. We want to be able to generate high-fidelity virtual humans instantly without any manual labor, develop deployable and low-cost capture systems that can be accessible by anyone, and enable real-time capabilities wherever possible. Immersive communication, AI, and virtual reality…these are the major applications that are pushing our technology.”

“Hao Li is a true creative force in vision and graphics worldwide” said Yannis C. Yortsos, dean of the USC Viterbi School of Engineering. “His new leadership role at ICT will help further cement its pioneering role in shaping the frontiers of virtual reality and will continue grow the impact of USC Computer Science in the vibrant Silicon Beach ecosystem.”

The German born Dr. Li obtained his PhD from ETH Zurich in 2010 and received his MSc degree in Computer Science in 2006 from the University of Karlsruhe (TH). He was a postdoctoral researcher at Columbia and Princeton Universities, a visiting professor at Weta Digital in 2014 and a visiting researcher at EPFL, Stanford University, National University of Singapore, and ENSIMAG.  Li is the CEO of Pinscreen. He received the SNF Fellowship for prospective researchers, and was named as part of the NextGen 10: Innovators under 40 by CSQ in 2014.

* Front page image Credit: LA Times.


Thanks so much for reading our article.

We've been a free service since 1999 and now rely on the generous contributions of readers like you. If you'd like to help support our work, please join the hundreds of others and become an fxinsider member.