Computer Scientist Yang Wang Awarded for Work on Computer Vision, Deep Learning
Computer Sciences Associate Professor Yang Wang was named the inaugural Faculty of Science Research Chair in Fundamental Science. The U of M researcher says the award ($20,000/year for three years) means he has more flexibility studying computer vision, machine learning and deep learning.
“The award has given me freedom to explore more fundamental research projects.”
Computer vision (CV), an interdisciplinary scientific field, focuses on how computers gain high-level understanding from digital images or videos. In layperson’s terms, CV is about developing techniques to help computers “see”.
Although the challenge of creating computer vision might seem simple, it remains unsolved for a variety of reasons. For example, it’s very hard for computer vision to interpret poses. It’s also difficult to distinguish people in a crowded scene.
For many computer vision applications, such as surveillance or traffic monitoring, people are the main interest. Therefore those kinds of applications need an algorithm that can accurately detect people as well as understand their poses and actions.
One of Wang’s projects is developing a vision algorithm that counts the number of people in an image. Another project is at the intersection of computer vision and natural language. He calls that ‘referring image segmentation’, which is important to intelligent image editing.
“The idea is that I give you an image and a sentence describing something in it. For example, ‘the man in black shirt next to the red car, in front of the building’. The algorithm identifies the mentioned objects.”
Wang and his team are also working on video captioning, where the algorithm “sees” a certain frame in a video and then generates a sentence describing what comes next. This applies to assistive technology for visually impaired people. By embedding this capability into a device, the device can warn of a potentially dangerous situation.For Wang, the study of computer science was unexpected. He grew up in Zhengzhou, the capital city of the Chinese province of Henan. His mother was a factory worker mother and his father a railway worker. No one in his family had attended college.
Wang’s first exposure to computers occurred when he was in primary school. He credits his high school programming teacher for inspiring him in the early stages of his computer science career.
“I got pretty serious in high school. I competed in a programming contest. So when I was choosing a major in college, I decided there was no other choice, I have to study computers.”
When he was a graduate student, his PhD advisor at Simon Fraser University (Dr. Greg Mori) had the biggest influence on Wang’s academic career.
“I saw how he worked, how he interacted with students, and how he organized his life. That’s how I got fascinated with an academic life and eventually decided to become a professor.”
Wang started at the UofM in 2012. He divides his time between teaching, advising and parenting his young son and daughter. His enjoys working with students on research projects.
“It’s just seeing students starting relatively green and growing into someone who can take on whatever challenges they face. That’s really my favourite part.”
For those interested in learning more about machine learning, Dr. Wang will deliver a talk Thursday, October 10, at 9:00am, on the topic of “Deep Learning for Video Summarization“, in EITC E2 125 at University of Manitoba, Fort Garry campus.
Who: Machine Learning Special Interest Group Research PresentationWhat: Presentation Title: Deep Learning for Video SummarizationSpeaker: Yang Wang, Associate Professor, Department of Computer ScienceWhen: Thursday October 10th, 2019Time: 9:00 am – 10:00 amLocation: Room : EITC E2 125 at Fort Garry campus
This special interest group is organized by Bill Leslie, Department of Internal Medicine, firstname.lastname@example.org and Lisa Lix, Department of Community Health Sciences, email@example.com and feel free to contact them if you are interested in actively participating in this special interest group.