Professor Jeong Je-chang's Team Earns the Top-score Achievement From AI-based Video Processing International Competition
Development of deep learning-based noise abatement technology
|Copy URL / Share SNS||
The team of Professor Jeong Je-chang of the Division of Electrical Engineering at Hanyang University achieved the highest-ranking performance at the NTIRE 2019 Academic Conference called the NTIRE 2019 Challenge on Real Image Denoising which was recently held in Long Beach, California, the U.S.
The challenge was co-hosted by CVPR 2019, which is the world's foremost academic conference in the field of computer vision sponsored by Samsung Electronics, Amazon, Adobe, and NVDIA, and by NTIRE 2019, which is an academic comference on new technologies for the restoration of images and the enhancement of picture quality. The challenge was a competition that effectively removes the actual noise contained in natural images and restores the images. It was conducted on two tracks: noise removal of camera sensor output images and noise removal of standard color images.
More than 200 researchers participated in each track and 15 teams, including South Korea, China, and the U.S., advanced to the final round. Jeong's team (researcher Park Beom-joon and Yoo Song-hyun) developed deep learning-based noise abatement technology and ranked second on each of the two tracks, earning the best overall score.
"We want to contribute to enhancing the competitiveness of the local image industry by developing core technologies by applying deep learning not only to image compression, but also to such areas as video restoration, image quality improvement and ultra-high resolution conversion," said Jeong. Jeong is the inventor of 70 patents registered as international standards in the field of imaging, including UHD TV's image compression technology HEVC.
Translated by Jeon Chae-yun firstname.lastname@example.org
This week's top news
Korean Traditional Colors
2017 JoongAng Ilbo University Rankings, Seoul Ranked 3rd · ERICA 9th
A 2018 Hanyang University Essay Entrance Exam Science & Engineering Departments
[Researcher of the Month] Hedging Deep Features for Visual Tracking
2019 Admission Requirements and Consultations Briefing Session
[Excellent R&D] BK21 PLUS, Supporting and Fostering Talented Students
[Researcher of the Month] New Synthesis of Metal-metalloid Material with Improved Catalyst Efficiency and Durability
[Researcher of the Month] The High Mobility of Single-Crystal Nanowires Opens Potential for Future Displays
[채널H] 2016 Chuseok event for foreign students
[Excellent R&D] Manpower Cultivation for Industry Convergence of 5G and Unmanned Vehicles