Byeonghyun Pak

I am a first lieutenant in the Republic of Korea Army (ROKA), currently serving as a research officer at the Agency for Defense Development (ADD), the Korean counterpart to the U.S. DARPA. Previously, I received my bachelor's degree from School of Transdisciplinary Studies at DGIST, where I was fortunate to work with Prof. Kyong Hwan Jin.

My work focuses on advancing Machine Learning (ML) and Computer Vision (CV). At ADD, I am contributing to the development of robust CV defense systems that perform reliably in adverse and out-of-distribution conditions.

profile photo

Email  /  CV (TBD)  /  Google Scholar  /  Github

Publications

Textual Query-Driven Mask Transformer for Domain Generalized Segmentation
Byeonghyun Pak*, Byeongju Woo*, Sunghwan Kim*, Dae-hwan Kim, Hoseong Kim
ECCV, 2024
project page / paper / code

We introduced a semantic segmentation method that generalizes well to unseen domains. Our method effectively handles domain shifts by using domain-invariant semantic knowledge from text embeddings of vision-language models (e.g., CLIP).

B-spline Texture Coefficients Estimator for Screen Content Image Super-Resolution
Byeonghyun Pak*, Jaewon Lee*, Kyong Hwan Jin
CVPR, 2023 (Highlight)
project page / paper / code

We proposed an implicit neural representation using 2D B-splines for high-quality screen content image (SCI) super-resolution. Our method aims to reconstruct the unique edges and textures of SCIs, minimizing information distortion and artifacts.


Credits to Jon Barron for this website's template.