Jaehyeok Bae

Hello! I am a first-year Ph.D. student in the Department of Electrical Engineering at Stanford University. I look forward to collaborate with professionals in the global research community, including both academia and industry.

I received my Bachelor's science (Summa Cum Laude) in Electrical and Computer Engineering at Seoul National University, where I was supported by Presidential Science Scholarship. Previously, I spent spent some time at Gauss Labs as a computer vision applied scientist intern.

Email  /  CV  /  Scholar  /  LinkedIn  /  Github

profile photo


I'm interested in computer vision, machine learning, image processing, along with biomedical imaging.

Specifically, my goal is to develop CV/ML applications for medical and industrial domains, alongside future medical devices, thereby significantly contributing to patient care, diagnosis, and treatment.

* denotes equal contribution.

clean-usnob PNI : Industrial anomaly detection using position and neighborhood information
Jaehyeok Bae*, Jaehan Lee*, Seyun Kim
ICCV, 2023
paper / video / poster / github

- Proposed a novel anomaly detection and localization alogrithm for industrial datasets, by training a normal feature distribution using position and neighborhood information of local features.

clean-usnob N-ImageNet: Towards robust, fine-grained object recognition with event cameras
Junho Kim, Jaehyeok Bae, Gangin Park, Dongsu Zhang, Youngmin Kim
ICCV, 2021
paper / video / github

- Introduced N-ImageNet, a large-scale dataset targeted for robust, fine-grained object recognition with event cameras.

- Empirically showed that pretraining on N-ImageNet improves the performance of event-based classifiers.

clean-usnob Design of a perforated panel for transmission noise reduction
Younghyo Park, Jaehyeok Bae, Jin Woo Lee
KSME, A, Vol. 39, No. 4, 2015
paper (in Korean)

- Proposed a design method for a perforated panel to reduce the level of incident noise without obstructing the flow of incoming fluid. (written in Korean)


clean-usnob SNU FastMRI Challenge
Jaehyeok Bae, Sungkyung Kim
Electrical and Computer Engineering, Seoul National University, 2022~2023
homepage / ppt / video (in Korean) / github

- Proposed an algorithm to restore aliased images from accelerated MRI scans into aliasing-free images, 2nd place award in the 2022 competiton.

- Served as the contest coordinator for the 2023 competition, evaluating and analyzing the participants' models.

Check out Jon Barron's repository for the template of this website.