WATERLOO — A University of Waterloo electrical engineer has won an Emmy Award for developing software that measures and improves the quality of video seen by millions of people every day.
Dr. Zhou Wang won the 67th Engineering Emmy Award for co-inventing what is called structural similarity — a software algorithm that quickly and accurately assesses how a video will be perceived by viewers.
That breakthrough in video engineering occurred 13 years ago at the University of Texas in Austin. Since then Wang has moved to the University of Waterloo, and co-founded a startup called SSIMWAVE Inc. that is commercializing further innovations in video engineering.
Engineering Emmy Awards are given out by the Television Academy for innovations that significantly improve the transmission, recording or reception of television.
Accuracy and simplicity made structural similarity a standard tool in broadcast and post-production houses throughout the television industry, says the Television Academy in a release.
It can be easily applied in real time on common processor software, and is now widely-used to test and refine video quality throughout the global cable and satellite TV industry. It directly affects the viewing experiences of tens of millions of viewers daily, the Television Academy says.
Wang did his PhD at the University of Texas in Austin under Dr. Allen Bovik in 1998. Wang and Bovik published their research on structural similarity in 2002. Bovik called Wang and told him about winning the Engineering Emmy Award. Wang did not even know about the nomination.
Wang’s PhD and post-doctoral work was on understanding how human brains process the information collected by eyes, optic nerves and the brain — the human visual system. Different parts of the visual system process different information from videos — movement, colour, contrast and lighting, among other things.
Wang developed computer algorithms that mimic the responses of the human-visual system. Those algorithms became known as the structural similarity. Before that, there was no way to easily measure the quality of video from a viewer’s perspective that was accurate and produced fast results.
By 2009, Wang was at the University of Waterloo and his paper on structural similarity had been cited 500 times. Wang supervised Kai Zeng and Abdul Rehman as they did their PhDs in electrical engineering. By the time they finished their doctorates, Wang’s seminal paper had been cited 10,000 times.
The feedback from Google, Intel, Qualcomm and Comcast convinced the trio to found the startup SSIMWAVE Inc. in 2013. It works out of the Accelerator Centre and the University of Waterloo. The startup has a suite of software tools that build on Wang’s work with structural similarity.
“Now we are presenting something that can be used by the industry to optimize, to save 50 percent on bandwidth, and to get live feedback on the video quality,” Rehman said.
It is an important technical achievement at a time when video streaming takes up a huge amount of the Internet’s capacity during peak hours for watching television.
“I think it was a very, very pleasant surprise,’ Rehman said of the Emmy Award during an interview in the startup’s laboratory.
Behind Rehman was a large-screen television that showed video of a downhill skier on the left side of the screen. On the right was the same video using SSIMWAVE Inc.’s technology. The difference in quality was stark — the one on the left broke up and pixilated, the one on the right was crisp and clear.
SSIMWAVE Inc. has six full-time employees and several part-timers. The quality index it developed manages and improves video displayed on movie screens, tablets, laptops, computers, and smartphones.
It can also be applied to medical imaging, Internet video, satellite images. It can be used to improve cameras and other equipment used to create videos. SSIMWAVE technology can also be used to predict video quality on future devices.
“You give me the properties of a display device that you are planning to launch in 2020 and I will tell you the perceived quality of an image or video compared to a TV that you have right now,” Rehman said.
The multimedia-entertainment-broadcast industry is SSIMWAVE’s customer base. Its biggest customer is Comcast, the largest broadcaster in the world.
It serves content producers like Sony, Fox and HBO.
“So we serve both, broadcasters and video-delivery providers like YouTube,” Rehman said.
The pioneering work Wang did 13 years ago on structural similarity means SSIMWAVE is well known and welcome around the world, Rehman said.
“A good thing about us is that we are very well recognized by the industry,” he added. “So everyone we talk to knows our contributions.”