Insoo Kim
Principal Researcher at Samsung Advanced Institute of Technology (SAIT)
Ph.D. Student at KAIST AI
insoo1.kim [AT] samsung.com
About
I am currently a Ph.D. Student of Algorithmic Intelligence Laboratory (ALIN-LAB) at KAIST AI, advised by Prof. Jinwoo Shin, where I focus on Motion Deblurring for real-world applications. In parallel, I work as a Principal Researcher at the Samsung Advanced Institute of Technology (SAIT), focusing on generative models for image generation and processing. Previously, I worked at Mando as a Senior Researcher, developing automotive radar signal processing algorithms for autonomous driving and ADAS systems. Before that, I received my M.S. degree in KAIST EE, advised by Prof. Hoirin Kim, where I conducted my research in Statistical Speech & Sound Computing Lab (SSSC LAB), focusing on speech-based emotion recognition.
Experience
Principal ResearcherNov. 2017 - Present
Samsung Advanced Institute of Technology (SAIT)
Senior ResearcherJan. 2011 - Oct. 2017
Education
KAISTAug. 2023 - Present
Ph.D Student, Graduate School of AI
KAISTFeb. 2009 - Feb. 2011
M.S., Graduate School of Electrical Engineering
Inha UniversityFeb. 2006 - Feb. 2009
B.S., Electronic Engineering
Publications
SpatialBoost: Enhancing Visual Representation through Language-Guided Reasoning
Byungwoo Jeon*, Dongyoung Kim*, Huiwon Jang, Insoo Kim, and Jinwoo Shin
NeurIPS Workshop 2025
RefDeblur: Blind Motion Deblurring with Self-Generated Reference Image
Insoo Kim, Geonseok Seo, Hyong-Euk Lee, and Jinwoo Shin
TMLR 2025
Controllable Blur Data Augmentation Using 3D-Aware Motion Estimation
Insoo Kim, Hana Lee, Hyong-Euk Lee, and Jinwoo Shin
ICLR 2025
Real-World Efficient Blind Motion Deblurring via Blur Pixel Discretization
Insoo Kim, Jae Seok Choi, Geonseok Seo, Kinam Kwon, Jinwoo Shin, and Hyong-Euk Lee
CVPR 2024
Quality-Agnostic Image Recognition via Invertible Decoder
Insoo Kim, Seungju Han, Ji-won Baek, Seong-Jin Park, Jae-Joon Han, and Jinwoo Shin
CVPR 2021
DiscFace: Minimum Discrepancy Learning for Deep Face Recognition
Insoo Kim, Seungju Han, Seong-Jin Park, Ji-won Baek, Jinwoo Shin, Jae-Joon Han, and Changkyu Choi
ACCV 2020
Meta Variance Transfer: Learning to Augment from the Others
Seong-Jin Park, Seungju Han, Ji-Won Baek, Insoo Kim, Juhwan Song, Hae Beom Lee, Jae-Joon Han, and Sung Ju Hwang
ICML 2020
Deep Speaker Representation Using Orthogonal Decomposition and Recombination for Speaker Verification
Insoo Kim, Kyuhong Kim, Jiwhan Kim, and Changkyu Choi
ICASSP 2019
Analysis on Mutual Interference in Automotive Radar System Using FMCW Waveform
Insoo Kim, Kyeongjin Song, Hwanjoon Kang, Seongung Yun, Seonghee Jeong, and Junnam Oh
AVEC 2012
Emotion Recognition Using Normalized Histogram of Speech Features
Insoo Kim, Myung Jong Kim, and Hoirin Kim
APSIPA 2010
Photorealistic Motion Deblurring via Flow-Based Generative Models
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
SpatialBoost: Enhancing Visual Representation through Language-Guided Reasoning
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
RefDeblur: Blind Motion Deblurring with Self-Generated Reference Image
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Controllable Blur Data Augmentation Using 3D-Aware Motion Estimation
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Real-World Efficient Blind Motion Deblurring via Blur Pixel Discretization
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Quality-Agnostic Image Recognition via Invertible Decoder
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
DiscFace: Minimum Discrepancy Learning for Deep Face Recognition
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Meta Variance Transfer: Learning to Augment from the Others
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Deep Speaker Representation Using Orthogonal Decomposition and Recombination for Speaker Verification
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Analysis on Mutual Interference in Automotive Radar System Using FMCW Waveform
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Emotion Recognition Using Normalized Histogram of Speech Features
Seong Hyeon Park, Huiwon Jang, Byungwoo Jeon, Sukmin Yun, Paul Hongsuck Seo, and Jinwoo Shin
Conference on Neural Information Processing Systems (NeurIPS), 2024,
Spotlight
Academic Services
Conference Reviewer: CVPR'23-26; ICCV'23,25; ECCV'24; ICLR'25; NeurIPS'25