Junseong Shin

Ph.D. Student · Hanyang University, VILAB

I am a Ph.D. student at VILAB, advised by Prof. Tae Hyun Kim. I build generative priors — diffusion and flow matching — to solve inverse problems in vision. My work spans image restoration (reflection removal, dehazing, shadow removal) and is expanding toward video generation and 3D scene understanding.

I received my B.S. in Mathematics from Hanyang University, and I now leverage this background in Bayesian methods and generative models for vision.

I am on the job market for research scientist and postdoctoral positions starting in 2027.

CV Email Scholar GitHub LinkedIn

News

Sep 2026Two papers are accepted to NeurIPS 2026!
Sep 2026Selected for the NRF Doctoral Student Research Grant as principal investigator!
Jun 2026One paper is accepted to ECCV 2026!
Jul 2025One paper is accepted to ICCV 2025!
Oct 2023One paper is accepted to CIKM 2023!

Publications

* equal contribution

PickMoment: Continuous-Time Single-Image-to-Video via Learning Deblurring and Blur-to-Video

NeurIPS 2026

PickMoment: Continuous-Time Single-Image-to-Video via Learning Deblurring and Blur-to-Video

Junseong Shin, Hyeonsu Jo, Daehyun Kim, Tae Hyun Kim

Paper Project Page

Image-to-Video · Deblurring · Continuous-Time · Video Generation

ODDR: One-Step Deshadow Diffusion via Reward Guidance

NeurIPS 2026

ODDR: One-Step Deshadow Diffusion via Reward Guidance

Junseong Shin, Kijun Kim, Minseong Kim, Dongjin Kim, Tae Hyun Kim

Paper Project Page

Shadow Removal · Diffusion · One-Step · Reward Guidance

PRISM: Latent Composition Consistency for Single-Image Reflection Removal

ECCV 2026

PRISM: Latent Composition Consistency for Single-Image Reflection Removal

Junseong Shin, Tae Hyun Kim

Paper Code Project Page

Reflection Removal · Flow Matching · Latent Space · Contrastive Learning

Revisit Haze Physical Model as ODE and Non-Homogeneous Haze Generation for Real-World Dehazing

ICCV 2025

Revisit Haze Physical Model as ODE and Non-Homogeneous Haze Generation for Real-World Dehazing

Junseong Shin*, Seungwoo Chung*, Yunjeong Yang, Tae Hyun Kim

Paper Code Project Page

Dehazing · Rectified Flow · Markov Chain · Brownian Motion

PAPP: Plug-And-Play Prompts for Shadow Removal

Under review

PAPP: Plug-And-Play Prompts for Shadow Removal

Junseong Shin*, Seungwoo Chung*, Tae Hyun Kim

Shadow Removal · Prompt Learning · Plug-and-play

Deep Variational Bayesian Modeling of Haze Degradation Process

CIKM 2023

Deep Variational Bayesian Modeling of Haze Degradation Process

Eun Woo Im*, Junseong Shin*, Sungyong Baik, Tae Hyun Kim

Paper Code

Dehazing · Variational Inference · Bayesian · Plug-and-play

Contact

Email is the best way to reach me: junsung6140@hanyang.ac.kr