Ten billion times a day people upload a video to Snapchat. From supermodels to cyber-celebrities to middle schoolers at the mall, people are using video to share personal stories on the social media app. While video sharing encourages company growth, storage can create challenges.
Zhu Li of the UMKC’s School of Engineering received a $10,000 grant from Snap, Inc., Snapchat’s developer, to support his research on video compression in order to address storage issues. Li has been working on the technology for the last 15 years — since the early days of his career with Motorola—but the need for the technology continues to grow.
“We are increasingly sharing videos, teenagers especially,” says Li. “It’s free for the user, but it’s not free for companies like Snap. The majority of the cost for them is to host and distribute video. So if we can shrink the storage 10 percent, then we shrink their costs 10 percent.”
Snap Inc., which defines itself as “a camera company,” is largely a video and image sharing platform. Jianchao Yang, Research Scientist at Snap, Inc., notes that video encoding is critical to the company.
“I’ve known Professor Zhu for a long time,” Yang says. He has extensive expertise in both academia and the industry. Snap deals with billions of videos every day. Better encoding can help our product in many aspects.”
Li’s research is extensive. He has 25 issued or pending patents and has contributed to more than 90 publications in book chapters, journals and conference proceedings. He notes that the subject of video compression still energizes him.
“In the technology field, video has shown some of the greatest progress.If you drove a car 10 years ago and you drive a car now, it’s pretty much the same. Four wheels, four doors. You’re basically driving at the same speed.”
Video technology has advanced exponentially.
“For video, 10 years ago you would have this tiny, grainy screen and a manageable, but relatively large, camera. The technology keeps getting better and better. Now we have 60-inch screens where the video quality is sharper and better. The camera might be your phone. The next step is even more interesting.”
The significance of Li’s research is not only working within the framework of compressing video as it exists now, though that is part of it, but managing the video technology that is coming.
“Once we can easily use a camera that takes a 360 degree picture, it will be possible to take a bird’s eye view. This will expand our vision exponentially.”
The experience of this will be drastically different than what we experience today.
“You will be virtually there,” says Li. “You will be living people’s experiences in real time.”
While the capability to create and store this type of video has an impact on companies like Snap, Inc., there are broader implications than social media sharing.
“You can virtually be sitting in any classroom, interacting with your peers, asking questions, but it will feel as if you’re in the room,” Li says. “It could impact remote medical treatment and surgeries.There are significant military applications. All the electronic industry is moving in this direction.”
But, Li pointed out, expanded virtual reality will increase the storage necessary to accommodate the resulting video. Currently, the structure of the internet cannot carry this volume of information. Compression of pixels is essential to ensure that video will be small enough to be carried.
“It’s difficult for this generation to understand how much the technology has changed,” says Li. “DVD was the first success of digital compression. Before that we had VHS tape. We used to carry a separate camera and calculator. The Walkman was a big deal. Now our phones do all of this.”
Compressing video is the answer to meeting increasing demand.
“Most people are focusing on recognition. Our approach is harder,” Li says. “We need to predict what will happen in the next frame. But we are making good progress.”