Jun-Hyeok Seo

I am a first-year Ph.D. student in the Department of Artificial Intelligence at Kyungpook National University, advised by Prof. Dong-Gyu Lee at the KNU Computer Vision & Machine Intelligence Lab.

Prior to KNU, I received my B.S. degree from the Department of Computer Engineering at Kumoh National Institute of Technology, advised by Prof. Hyeonju Yoon. My research focuses on human-centric AI and computer vision.

News

2026

May 2026One paper is under review for Conference on Neural Information Processing Systems (NeurIPS).
Jan 2026One paper is under review for IEEE Journal of Biomedical and Health Informatics (JBHI).

2025

Oct 2025Received the Top Prize from the AI Convergence Safety Idea Competition by KNU.
Sep 2025I was selected as a Full Scholarship Awardee for my Ph.D. program at KNU.
Sep 2025One paper was presented at Korea Artificial Intelligence Conference (KAIC).
Aug 2025Received the Encouragement Prize from the AI Hackathon Competition by CNU.
Jul 2025One paper was presented at IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRw).
Jun 2025Received the 3rd Place in the VizWiz Zero-Shot Image Classification Grand Challenge at CVPR Workshops 2025.
Apr 2025One paper was presented at KIISE Transactions on Computing Practices (KTCP).

2024

Dec 2024Received the Best Presentation Paper Award at Korea Software Congress (KSC).
Sep 2024Received the Merit Scholarship (IH003) from KNU.
Jun 2024Received The Mayor of Daegu Metropolitan City Award at the Daegu Public Data Startup Competition.
Mar 2024Joined Computer Vision and Machine Intelligence Lab. at Kyungpook National University.
Mar 2024Received the Merit Scholarship (IH003) from KNU.

Publications

* indicates equal contribution. † indicates the corresponding author.

International Publications

NeurIPS 2026

DeACT: Disentangling Action and Scene via Cross-Attentive Purification for Explainable Action Recognition

Jun-Hyeok Seo, Dong-Gyu Lee†

Conference on Neural Information Processing Systems (NeurIPS), 2026

PaperUnder Review

This work is currently under review.

JBHI 2026

NiRA: Noise-invariant Causal Discrete Representation Alignment for Video-based Remote Physiological Measurement

Jun-Hyeok Seo, Da-Hee Kim, Jeong-Cheol Lee, Seungyeon Lee, Sangtae Ahn, Dong-Gyu Lee†

IEEE Journal of Biomedical and Health Informatics (JBHI, IF: 6.7), 2026

PaperUnder Review

This work investigates noise-invariant causal discrete representation alignment for robust video-based remote physiological measurement.

CVPRw 2025

Multi-Perspective LVLM Prompting for Robust Zero-Shot Image Classification

Wonjun Choi, Jeong-Cheol Lee, Jun-Hyeok Seo, Dong-Gyu Lee†

IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRw), 2025

Paper

This paper proposes a multi-perspective prompt-based framework for zero-shot image classification using large vision-language models.

Domestic Publications

KAIC 2025

Context-Aware Adaptive Mask Selection for Consistent Referring Video Object Segmentation

Jun-Hyeok Seo, Young-Woo Youn, Wonjun Choi, Dong-Gyu Lee†

Korea Artificial Intelligence Conference (KAIC), 2025

Paper

This paper proposes a context-aware adaptive mask selection framework for referring video object segmentation. The method integrates CLIP-based keyframe selection, mask consistency scoring, text-guided object number determination, and temporal mask interpolation.

KTCP 2025

Enhancing Detail Quality in Latent Diffusion Model-Based Super-Resolution via Multi-Scale Discrete Wavelet Transform

Jun-Hyeok Seo, Dong-Gyu Lee†

KIISE Transactions on Computing Practices (KTCP), 2025

Paper

This paper proposes a multi-scale discrete wavelet transform framework to enhance detail quality in latent diffusion model-based super-resolution.

KSC 2024

Diffusion Model Based Image Super-Resolution with Multi-Scale High Frequency Error Maps

Jun-Hyeok Seo, Dong-Gyu Lee†

Korea Software Congress (KSC), 2024

PaperBest Presentation Paper Award

This paper proposes a diffusion model-based image super-resolution framework using multi-scale high-frequency error maps for accurate restoration of image details.

KIIT 2023

Robust Captcha Image Recognition Algorithm Using Transformer-Based OCR Model

Jun-Hyeok Seo, Seongjin Jang, Minjun Kim, Hyeonju Yoon†

Proceedings of the Korean Institute of Information Technology Conference (KIIT), 2023

PaperBronze Award

This paper proposes a robust OCR algorithm using an end-to-end Transformer-based TrOCR model for captcha image recognition.

KCC 2023

Deep Learning-Based Vehicle Damage Area Detection System using Semantic Segmentation

Jun-Hyeok Seo, Hyeonju Yoon†

Korea Computer Congress (KCC), 2023

Paper

This paper proposes a vehicle damage detection system through semantic segmentation using UNet++.

KIIT 2023

Segmentation of Stain Area in Clothing Images using UNet++

Jun-Hyeok Seo, Gunwook Kim, Junsu Park, Sungyoung Kim†

Proceedings of the Korean Institute of Information Technology Conference (KIIT), 2023

PaperBronze Award

This paper proposes a deep learning-based technique for detecting stains in clothing images using UNet++.

KIIT 2022

Domain Dictionary Construction for Automatic Analysis of Subject CQI Reports

Gunwoo Kim, Junsu Park, Jun-Hyeok Seo, Yuchul Jung, Hyeonju Yoon†

Proceedings of the Korean Institute of Information Technology Conference (KIIT), 2022

PaperSilver Award

This paper constructs a domain dictionary to support automatic analysis of subject CQI reports.

KIIT 2022

Deep Learning-Based Vehicle Model and License Plate Identification System using Vehicle Image

Jun-Hyeok Seo, Hyeonju Yoon†

Proceedings of the Korean Institute of Information Technology Conference (KIIT), 2022

PaperGold Award

This paper proposes a vehicle model recognition and license plate identification system using YOLOv5.

Education

Integrated M.S.-Ph.D. in Artificial Intelligence

Kyungpook National University

Mar. 2024 - Present | Daegu, Korea

Integrated M.S.-Ph.D. in AI Safety Convergence (Minor)

Kyungpook National University

Mar. 2024 - Present | Daegu, Korea

B.S. in Computer Engineering

Kumoh National Institute of Technology

Mar. 2020 - Feb. 2024 | Gumi, Korea

Experience

Technology Transfer

Happyaging

Jul. 2025 - Dec. 2025 | Daegu, Korea

Undergraduate Tutor

Kyungpook National University

Mar. 2025 - Jul. 2025 | Daegu, Korea

Education Completion

LG Aimers

Jul. 2022 - Sep. 2022 | Online

Web Publishing Team

TimeComs

Jul. 2022 - Sep. 2022 | Gumi, Korea

Facilities Management Department - Data Processing

WONIK QnC

Jul. 2021 - Sep. 2021 | Gumi, Korea

Freshmen Mentor

Kumoh National Institute of Technology

Mar. 2021 - Feb. 2022 | Gumi, Korea

Achievements

Scholarships

Doctoral Program Full Scholarship from KNU, 2025-2
Merit Scholarship (IH003) from KNU, 2024-2
Merit Scholarship (IH003) from KNU, 2024-1

Awards

Top Prize, AI Convergence Safety Idea Competition by KNU, 2025
3rd Place Winner, VizWiz Zero-Shot Image Classification Grand Challenge in CVPR 2025, 2025
Encouragement Prize, AI Hackathon Competition by CNU, 2025
The Mayor of Daegu Metropolitan City Award, Daegu Public Data Startup Competition, 2024
Best Presentation Paper Award, Korea Software Congress, 2024
Bronze Award for Outstanding Paper, Proceedings of KIIT Conference, 2023
Bronze Award for Outstanding Paper, Proceedings of KIIT Conference, 2023
Silver Award for Outstanding Paper, Proceedings of KIIT Conference, 2022
Gold Award for Outstanding Paper, Proceedings of KIIT Conference, 2022

Teaching Assistant

Data Analysis (CAIB226), Undergraduate, 2025-1
Problem-Solving Skills and Algorithms (CAIB0230), Undergraduate, 2024-2