The best paper award at Korea Computer Graphics Society Conference 2016
The paper presented by Gyuhyun Lee (3rd year UNIST CSE PhD student) and co-authored by Tran Minh Quan (5th year UNIST CSE PhD student) and Prof. Won-Ki Jeong is selected as the best paper at Korea Computer Graphics Society (KCGS) Conference 2016.
This paper proposes a novel dictionary learning approach for biomedical image segmentation. The proposed algorithm employs dual dictionary for encoding features from example images and label images independently with a common sparse map to link both dictionaries. The core technique used in this work is convolutional sparse coding, which is a state-of-the-art shift-invariant dictionary learning approach. The proposed method is faster than deep learning while providing a similar accuracy without a time-consuming training time.
KCGS conference is the oldest and largest computer graphics conference in Korea, and the best paper award is the highest honor given to the best work presented at the conference.