[The Lab is led by professors Suha Kwak, Jaesik Park and Minsu Cho.]

The Computer Vision Lab at POSTECH will present the largest number of papers among Korean institutes participating at the 2022 Conference on Computer Vision and Pattern Recognition (CVPR).

The CVPR conference, which has been co-hosted by the Institute of Electrical and Electronics Engineers (IEEE) and the Computer Vision Foundation (CVF) since 1983, is the most influential international academic conference in AI. The Google Scholar h-index ranks it as high as 4th in the science and engineering sector, indicating its great influence (cf. Nature rank 1st, Science ranks 3rd and Cell ranks 7th). It is also mentioned as one of the three major academic conferences in computer vision along with the International Conference on Computer Vision (ICCV) and the European Conference on Computer Vision (ECCV).

The lab, led by Professor Minsu Cho (Department of Computer Science and Engineering, Department of Convergence IT Engineering) and professors Jaesik Park and Suha Kwak (Department of Computer Science and Engineering, Graduate School of Artificial Intelligence), will present 16 papers which is the largest number of papers among the participating Korean university labs and corporate research institutes this year.

It is globally rare to see this many papers being accepted from a single lab and the number of papers from each professor is the largest number in Korea as well. Moreover, it is noteworthy that the professors from different specialties in the field of computer vision worked closely together in a lab to reach this achievement.

Professor Minsu Cho is an expert in image registration. He focuses on the artificial intelligence that learns and reasons the associations among various visual elements in video, such as symmetry detection, action recognition, and geometric shape assembly.

Professor Jaesik Park excels in the field of 3D vision, which deals with reconstruction and recognition of 3D spaces. He has proposed a technique for inferring 3D spatial layouts from an image of an object – a technique for understanding 3D space using deep learning – and an AI painting technique for the metaverse space.

Professor Suha Park achieved stellar results in training video-recognition AI with minimum information. He also studies AI that deals with both video and natural language, and an AI technology that can perform strongly in a variety of unfavorable conditions.

The CVPR 2022 will be held both online and offline from June 19 to 24, 2022 (local time) in New Orleans, Louisiana.