Byeongjun Park

Ph.D. Student
KAIST
pbj3810@kaist.ac.kr

Education

[Mar. 2020 ~ Aug. 2025 (expected)] : Candidate for Integrated M.S. / Ph.D. degree, School of Electrical Engineering @ KAIST.
[Mar. 2015 ~ Feb. 2020] : B.S. degree (Cum Laude), School of Electrical Engineering @ KAIST.
[Mar. 2013 ~ Feb. 2015] : Gyeongnam Science High School.

Work Experience

[Mar. 2018 ~ Aug. 2018]: AI Research Intern @ Koh Young Technology

Research Interests

Generative Models: Diffusion Models, Novel View Synthesis, Neural Rendering

News

[Jul. 2024] One paper about Diffusion Model is accepted to ECCV 2024.
[Mar. 2024] One paper about Novel View Synthesis is accepted to TPAMI 2024.
[Feb. 2024] One paper about One-Image-to-3D is accepted to CVPR 2024.
[Jan. 2024] One paper about Diffusion Model is accepted to ICLR 2024.
[Oct. 2023] One paper about Dynamic Radiance Fields is accepted to WACV 2024.

Publications

* indicates equal contribution

Journal

ArXiv

DiffRef3D: A Diffusion-based Proposal Refinement Framework for 3D Object Detection

Se-Ho Kim*, Inyong Koo*, Inyoung Lee, Byeongjun Park, Changick Kim

Under Review in IEEE Transactions on Image Processing (TIP)

PDF

TPAMI

Bridging Implicit and Explicit Geometric Transformation for Single-Image View Synthesis

Byeongjun Park*, Hyojun Go*, Changick Kim

IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2024

PDF

Conference

arXiv

Diffusion Model Patching via Mixture-of-Prompts

Seokil Ham*, Sangmin Woo*, Jin-Young Kim, Hyojun Go, Byeongjun Park, Changick Kim

Preprint

PDF Code Project Page

ECCV

Switch Diffusion Transformer: Synergizing Denoising Tasks with Sparse Mixture-of-Experts

Byeongjun Park, Hyojun Go, Jin-Young Kim, Sangmin Woo, Seokil Ham, Changick Kim

European Conference on Computer Vision (ECCV), 2024

PDF Code Project Page

CVPR

HarmonyView: Harmonizing Consistency and Diversity in One-Image-to-3D

Sangmin Woo*, Byeongjun Park*, Hyojun Go, Jin-Young Kim, Changick Kim

The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024

PDF Code Project Page Demo

ICLR

Denoising Task Routing for Diffusion Models

Byeongjun Park*, Sangmin Woo*, Hyojun Go*, Jin-Young Kim*, Changick Kim

International Conference on Learning Representations (ICLR), 2024.

PDF OpenReview Code Project Page

WACV

Point-DynRF: Point-based Dynamic Radiance Fields from a Monocular Video

Byeongjun Park, Changick Kim

IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024.

PDF

ICIP

Temporal Flow Mask Attention for Open-Set Long-Tailed Recognition of Wild Animals in Camera-Trap Images

Jeongsoo Kim, Sangmin Woo, Byeongjun Park, Changick Kim

IEEE International Conference on Image Processing (ICIP), 2022.

PDF

ICIP

Fine-Grained Multi-Class Object Counting

Hyojun Go, Junyoung Byun, Byeongjun Park, Myung-Ae Choi, Seunghwa Yoo, Changick Kim

IEEE International Conference on Image Processing (ICIP), 2021.

PDF Code

Patents

[Sep. 2022] Method and apparatus with scene flow estimation (US20220301190A1) [link]

Projects

[Jul. 2024 ~ Present]: LIG Nex1 - Underwater sonar image enhancement with AI
[Apr. 2023 ~ Present]: SOOMVI - Precise landing technology for drones
[Sep. 2020 ~ Feb. 2022]: Ministry of Science and ICT (MSIT, Korea) - Digital ecology with AI
[Aug. 2020 ~ Aug. 2021]: Samsung Advanced Institude of Technology (SAIT) - Monocular scene flow estimation

Academic Services

Invited Talks
- [Aug. 2024] ETRI - Recent Trends in 3D Content Creation
Conference Reviewer
- ICLR @ 2025
- NeurIPS @ 2024
- ECCV @ 2024
- AAAI @ 2021, 2022
Journal Reviewer
- IET Computer Vision @ 2023

Awards & Honors

Samsung Best Paper Award in IEIE Autumn Annual Conference ($ 1,000) @ Nov. 2021
- Jeongsoo Kim, Byeongjun Park, Seungju Cho, Myung-ae Choi, Jinyoung Park, Hyungsoo Seo, Changick Kim, “An Optical Flow Fusion Deep Learning Network using Trap Camera Image Dataset”
National Academic Excellence Scholarship (Science and Engineering, Korea) @ 2016, 2017, 2018

Powered by Jekyll and Minimal Light theme.