Video Footage Helps Predict the Future

Machine learning meets security to protect the nation

There are about 30,000 domestic flights per day in the US on average; this means that security measures for detecting suspicious behavior are not only necessary but expensive. Imagine if computers were able to detect criminal activity before it happened more accurately than trained personnel. Research scientists are working to pair security, multimedia data, and machine learning to provide novel tools that can do just that. Dr. Yun Raymond Fu, of Northeastern University, is at the forefront of research focused on machine learning, computer vision, data mining, and social media analytics. Driven by the explosion of diverse multimedia from the Internet including personal or publicly available photos and videos, his team brings fundamental theory to synergetic media learning systems that collect data and is able to interact with precise analysis and possible suggestions.

Dr. Fu’s expertise allows him and his collaborative team to make early predictions about events and activities given the social context and billions of images or data points. Thus, the team approaches visual problems by asking, “what is it?” but even more importantly, “what should we do?” Thus, his research helps to mitigate the uncertainties that are more challenging for human personnel and allow for applications to operate at their optimal performance levels. His research is making significant impacts on areas of national defense in that it provides a solid solution in developing situational awareness. Additionally, his work has applications in improving therapeutic opportunities for patients that are disabled, increasing social connection using social media data, and making accurate predictions using video footage. In short, Dr Fu's research is a creative solution to how we can use the vast amount of data available to increase safety and well-being for our entire nation.

Current research includes:

  • Rehab at Home: Using his expertise in addition to collaborations with experts in the exercise field, Dr. Fu develops rehabilitation tools for patients to conduct therapy remotely in their own homes. By breaking down the barrier of care that requires patients to go to the hospital for rehab, Dr. Fu’s cell phone or kinect sensor based monitoring system is helping to improve quality of care, especially for patients with knee arthritis who otherwise would have difficulties leaving the home.

  • Predicting the Future: Ordinarily, security systems need to be monitored by someone that is trained to detect if something suspicious might arise in criminal activity. Dr. Fu has created a tool that can read and predict what will happen in the future given video input. Thus, his technology is able to objectively analyze footage more efficiently and effectively than even trained human personnel. Dr. Fu’s tool will have wide ranging applications including for security at the airport or battle field in addition to the predictions of everyday life for example, who might win a sporting event based on the first thirty minutes of play.

  • Therapeutic Detection: Dr. Fu built a framework that could facilitate a therapist to help understand what is abnormal using video footage of a patient to detect abnormal behavior more quickly and accurately. Using a deep neural network structure to facilitate the work, Dr. Fu’s research is guided by neurological feedback necessary for an objective analysis.

  • Enhancing Connections: Dr. Fu is looking at data collected by social media networking sites to develop better recommendations for potential networking or friendship connections. Currently a user on Facebook has not met 60-80% of their “friends”; Dr. Fu hopes that by further enhancing the ability to recommend more likely friendships, he and his team will help facilitate closer social ties.

  • Pinpointing Location: Dr. Fu is hoping to combine his intelligent detection technologies with geolocation. Using only data from a photograph, he and his team want to be able to detect the location, occupation, and demographic information of the people and setting of the photo. Dr. Fu’s tool would be an important security measure for recognizing where terrorists are or other criminals in addition to finding missing persons and more.

Bio

Dr. Fu still remembers the first time he saw a personal computer at a friend’s home while he was still in middle school. Up to that point, he had never seen a machine that was as powerful or had the potential to make as great an impact. From that time on, Dr. Fu believed that computing would be an important piece to changing people’s lives. Therefore, he began trying to learn all he could about the computing field, even at his young age.

Later, during his B.Eng. study, Dr. Fu was selected in a special program for gifted students. This gave him the opportunity to take courses across multiple disciplines. Through this experience, he developed a greater interest in computing and artificial intelligence. This also solidified his belief that, “the more intelligent machines people make, the more people can understand about themselves.” Dr. Fu continues to be motivated by the promise that these advances will benefit all of society by transforming the research community and impacting the world.

Aside from research, in his free time, Dr. Fu enjoys playing and listening to music. As a young child, he was passionate about playing the clarinet and flute. He also enjoys playing basketball, tennis, and swimming or traveling with family.

Website: http://www1.ece.neu.edu/~yunfu/

Publications

Low-Rank and Sparse Modeling for Visual Analysis

PDF

Human-Centered Social Media Analytics

PDF

Graph Embedding for Pattern Classification

PDF

Age Synthesis and Estimation via Faces: A Survey

PDF

Prediction of Human Activity by Discovering Temporal Sequence Patterns

PDF

Awards

ONR Young Investigator Award, Office of Naval Research, 2014

ARO Young Investigator Award, Army Research Office, 2014

INNS Young Investigator Award, International Neural Networks Society, 2014

Google Faculty Research Award, Google Research, 2010

China Government Award for Outstanding Self-financed Students Abroad, China Government, 2007