Curriculum Vitae
Jeonghyeok Do
Ph.D. Candidate (4th yr.)
Korea Advanced Institute of Science and Technology (KAIST)
School of Electrical Engineering
Video and Image Computing Lab. (VIC Lab.)
Advisor: Prof. Munchurl Kim
E-mail: ehwjdgur0913@kaist.ac.kr
YouTube: @dosensei97
CV: [PDF]
Office: Room 1106, N24 LG Innovation Hall, KAIST
Research Interests
Deep Learning
Convolutional Neural Networks (CNN)
Generative Adversarial Networks (GAN)
Graph Convolutional Networks (GCN)
Low-level Computer Vision
Image Super Resolution
Image Dehazing
High-level Computer Vision
Video Anomaly Detection
Multi-modal Action Recognition
Skeleton-based Action Recognition
International Publications
MorphVAD: Efficient Video Anomaly Detection Using Morphological Transformation
Jeonghyeok Do and Munchurl Kim
International Conference on Visual Communications and Image Processing (VCIP), IEEE, Dec. 2023.
[PDF]
Multi-modal Transformer for Indoor Human Action Recognition
Jeonghyeok Do and Munchurl Kim
2022 22nd International Conference on Control, Automation and Systems (ICCAS). IEEE, pp. 1155-1160, Nov. 2022.
[PDF]
Pseudo-Supervised Learning for Semantic Multi-Style Transfer
Learning-based JND-directed HDR Video Preprocessing for Perceptually Lossless Compression with HEVC
NTIRE 2020 Challenge on NonHomogeneous Dehazing
Codruta O. Ancuti, Cosmin Ancuti, Florin-Alexandru Vasluianu, Radu Timofte [and 48 others, including Jeonghyeok Do]
In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), 2020.
[PDF]
Domestic Publications
Jeonghyeok Do, and Munchurl Kim, “Priors on Bone Length Consistency for Skeleton-based Action Recognition,” Korea Institute of Broadcast and Media Engineers (KIBME) Workshops, Jun. 2023.
Jeonghyeok Do, Hyeonjun Sim, Gayoung Lee, Seongho Baek, and Munchurl Kim, “Efficient Video Anomaly Detection via Future Frame Prediction and Dual Memory Network,” Korea Institute of Industrial Engineers (KIIE) Workshops, Nov. 2021.
Jeonghyeok Do, and Munchurl Kim, “Efficient Video Anomaly Detection with Future Frame Prediction,” Korea Institute of Military Science and Technology (KIMST) Workshops, Nov. 2021.
Jeonghyeok Do, and Munchurl Kim, “Wafer Map Defect Pattern Classification with Progressive Pseudo-Labeling Balancer,” Korea Institute of Broadcast and Media Engineers (KIBME) Workshops, Nov. 2020.
Jeonghyeok Do, and Munchurl Kim, “Adaptive Attention based U-Net for Image Dehazing,” Korea Institute of Military Science and Technology (KIMST) Workshops, Nov. 2020.
Awards and Honors
KAIST President Award, Feb. 2015
Hanseong Nobel Scholarship Student ($5,000), Sep. 2014
Education
Ph.D. Student in Electrical Engineering, KAIST
Mar. 2021 - Present
Advisor: Prof. Munchurl Kim
G.P.A: 4.21 (40 credits completed, including those of M.S.)
M.S. in Electrical Engineering, KAIST
Mar. 2019 - Feb. 2021
Advisor: Prof. Munchurl Kim
G.P.A: 4.20 (51 credits completed)
B.S. in Electrical Engineering, KAIST
Mar. 2015 - Feb. 2019
G.P.A: 3.80 (151 credits completed)
Cum Laude
Double Major: Department of Mathematical Sciences
ChungBuk Science High School (CBSH)
Mar. 2013 - Feb. 2015
Early graduation (2 years)
Major Skills
Deep Learning Framework
PyTorch
Programming Languages
C, Matlab, Python
Operating System
Window
Document Work Tools
Word, PowerPoint, Excel, Visio, LaTeX, PyCharm
Project Experiences
LG Electronics Co., Ltd.
Low complexity deep Learning based super-resolution. (Support)
Sep. 2019 - Jan. 2020
SK Hynix
Defect Auto Define with Edge BMW Classification. (Project Lead)
Jan. 2020 - Nov. 2020
SK Hynix
Video Anomaly Detection based on Image Sequence Prediction. (Project Lead)
Jan. 2021 - Oct. 2021
Agency for Defense Development (ADD) - Center for Applied Research in Artificial Intelligence (CARAI)
MX-24: Soldier/Unit Action Recognition based on Multi-modal Learning. (Project Lead)
Jan. 2020 - Dec. 2022
Workshops
Workshop on Deep Learning Technology and Real-Time Implementation of Deep Technology
Topic : Deep Learning, Super Resolution, Frame Interpolation, HDR, Neural Networks Compression, GPU
Host Institution : Open Standards and ICT Association (OSIA)
Aug. 2019