Jaesung Rim

I received my Ph.D. from POSTECH, where I was advised by Prof. Sunghyun Cho. I received my Master's degree in ICE from DGIST and my bachelor's degree from Kwangwoon University.

My research interests include computational photography, particularly image restoration, image deblurring, low-light imaging, and mobile imaging.

Email / GitHub / Google Scholar / CV / LinkedIn

Datasets

RealBlur: The first real-world blur dataset for learning-based motion deblurring.
- RealBlur is now utilized as a standard benchmark.
RSBlur: The first blur dataset for analyzing the gap between real-world and synthetic blur.
- We organized AIM 2025 Challenge using RSBlur.
ExLPose: A human pose dataset with paired low-light and well-lit images.
HCBlur: A dual-camera blur dataset captured with iPhone 13 Pro Max.
GyroVD: The first gyro-based video deblurring dataset captured with GP 9 Pro XL.

Publications

	Gyro-based Deep Video Deblurring Jaesung Rim, Woohyeok Kim, Haeyun Lee, Heemin Yang, Ke Wang, Sunghyun Cho CVPR, 2026 Paper / Project / Code We proposed the first learning-based methods for gyro-based video deblurring. For training and evaluation, we collected the GyroVD dataset using GP 9 Pro XL.
	Dynamic Exposure Burst Image Restoration Woohyeok Kim, Jaesung Rim, Daeyeon Kim, Sunghyun Cho CVPR, 2026 Paper / Project / Code
	Towards Unified Image Deblurring using a Mixture-of-Experts Decoder Daniel Feijoo, Paula Garrido-Mellado, Jaesung Rim, Alvaro Garcia, Marcos V. Conde LoViF CVPRW, 2026 Paper / Code
	Gyro-based Neural Single Image Deblurring Heemin Yang, Jaesung Rim, Seungyong Lee, Seung-Hwan Baek, Sunghyun Cho CVPR, 2025 Paper / Project / Code
	Deep Hybrid Camera Deblurring for Smartphone Cameras Jaesung Rim, Junyong Lee, Heemin Yang, Sunghyun Cho SIGGRAPH, 2024 Paper / Project / Code We proposed a novel deblurring framework that simultaneously utilizes wide and ultra-wide cameras on smartphones. For training and evaluation, we collected the HCBlur dataset using iPhone 13 Pro Max.
	Burst Image Super-Resolution with Base Frame Selection Sanghyun Kim, Min Jung Lee, Woohyeok Kim, Deunsol Jung, Jaesung Rim, Sunghyun Cho, Minsu Cho (equal contribution) NTIRE CVPRW, 2024* Paper / Project
	ExBluRF: Efficient Radiance Fields for Extreme Motion Blurred Images Dongwoo Lee, Jeongtaek Oh, Jaesung Rim, Sunghyun Cho, Kyoung Mu Lee ICCV, 2023 Paper / Code
	Human Pose Estimation in Extremely Low-Light Conditions Sohyun Lee, Jaesung Rim, Boseung Jeong, Geonu Kim, ByungJu Woo, Haechan Lee, Sunghyun Cho, Suha Kwak (equal contribution) CVPR, 2023* Paper / Project / Code We proposed the ExLPose dataset, which provides pairs of well-lit and low-light images along with their pose labels. We presented a novel method utilizing the knowledge from both images.
	Realistic Blur Synthesis for Learning Image Deblurring Jaesung Rim, Geonung Kim, Jungeon Kim, Junyong Lee, Seungyong Lee, Sunghyun Cho ECCV, 2022 Paper / Project / Code Deblurring networks trained on synthetic datasets often struggle with real-world images. To better handle real-world blurred images, we analyzed the differences between real and synthetic blur on the RSBlur dataset and proposed a realistic blur synthesis pipeline that includes a camera ISP simulation.
	Iterative Filter Adaptive Network for Single Image Defocus Deblurring Junyong Lee, Hyeongseok Son, Jaesung Rim, Sunghyun Cho, Seungyong Lee CVPR, 2021 Paper / Project / Code
	Real-World Blur Dataset for Learning and Benchmarking Deblurring Algorithms Jaesung Rim, Haeyun Lee, Jucheol Won, Sunghyun Cho ECCV, 2020 Paper / Project / Code We proposed the RealBlur dataset, which is the first real-world blur dataset for learning-based methods. This dataset is now utilized as a standard benchmark.