Professor Sangmin Lee (School of Information Convergence) and his research team, including master's student Ji-yeon Kang, won the Outstanding Paper Award at the '2024 Spring Conference of the Korean Institute of Smart Media & Society for E-business Studies' for their presentation of the paper titled 'Measuring the Prediction Uncertainty of Human Pose Estimation Models for Personalized Rehabilitation Exercises.'

Ji-yeon Kang, a master's student from Professor Sangmin Lee’s research team (AI as a Service Laboratory) at Kwangwoon University, was awarded the Outstanding Paper Award at the '2024 Spring Conference of the Korean Smart Media Society & Korea Electronic Commerce Society' held at the Gwangju Realistic Content Cube (GCC) and Gwangju CGI Center on May 17-18.

The conference, attended by over 300 members of the Korean Media Society, was themed 'Experience the Future of Smart Media: Personalized, Interactive and Immersive,' and featured 156 paper presentations and 168 media art exhibits from 13 countries.

Ji-yeon Kang received the award for her paper on personalized rehabilitation exercises tailored to individual patient's physical conditions and disabilities, titled 'Measuring the Prediction Uncertainty of Human Pose Estimation Models for Personalized Rehabilitation Exercises.'

Recent studies on human pose estimation models focus on collecting relative coordinate information of joints to estimate the range and accuracy of joint movements. However, this technique is more suited to data from non-disabled individuals, and uncertainty in joint position estimation increases when parts of the body are obscured. Ji-yeon Kang’s research aims to reduce prediction uncertainty in human pose estimation and improve the accuracy of range estimation using directed acyclic graph neural networks, thus enhancing the accuracy of individual movement range estimations and establishing personalized exercise capability assessment criteria.

(From left) Kwangwoon University graduate students in AI Applications, Ji-yeon Kang, Hyo-young Jang, Seung-joon Baek, and Ho-joon Kim.