AFFILIATION

I am working at Aizawa-Yamasaki-Matsui Lab, as a Ph.D. candidate.

INTERESTS

My research interests are in deep learning for computer vision and natural language processing tasks. My primary interest is the interdisciplinary field between computer vision and natural language processing, and I have developed the scene text recognition model which is similar to image captioning. I am also interested in synthetic data generation to improve model performance.

PROFESSIONAL EXPERIENCE

• Researcher (alternative military service), Clova AI Research, NAVER/LINE Corp.
  • Jan. 2018 – Mar. 2020
  • Developed scene text recognition (STR) model which recognizes text in the natural scene.
    
    
• Researcher (alternative military service), Language Analytics, NCSOFT Corp.
  • Apr. 2016 – Dec. 2017
  • Developed sentence embedding model for question/document clustering and text style transfer model for colloquial text generation.

EDUCATION

• Ph.D. of Information Science and Technology, The University of Tokyo
  • Apr. 2020 - Present
  • Advisor: Prof. Kiyoharu Aizawa
    
    
• Master of Informatics, Kyoto University
  • Apr. 2014 – Mar. 2016
  • GPA: 3.9/4 (ABC scale) Rank is unknown.
  • Advisor: Prof. Shin Ishii and Kenji Doya (Adjunct Unit with KU lab & OIST lab)
  • Dissertation: Descriptive, generative, and hybrid approaches for neural connectivity inference from neural activity data.
    
    
• Bachelor of Mechanical System Engineering, Tokyo University of Agriculture and Technology
  • Apr. 2010 – Mar. 2014
  • GPA: 3.89/4 (ABC scale) Summa cum laude (GPA 1st/121)
  • Advisor: Prof. Ikuo Mizuuchi
  • Dissertation: Considerate Behavior of Robots based on Individual’s Preference

PUBLICATIONS

Computer vision top conference

Character Region Attention For Text Spotting
Youngmin Baek, Seung Shin, Jeonghun Baek, Sungrae Park, Junyeop Lee, Daehyun Nam, Hwalsuk Lee.
ECCV 2020.

CLEval: Character-Level Evaluation for Text Detection and Recognition Tasks
Youngmin Baek, Daehyun Nam, Sungrae Park, Junyeop Lee, Seung Shin, Jeonghun Baek, Chae Young Lee, Hwalsuk Lee.
Workshop on Text and Documents in the Deep Learning Era, CVPR WTDDLE 2020.
[Paper]

On Recognizing Texts of Arbitrary Shapes with 2D Self-Attention
Junyeop Lee, Sungrae Park, Jeonghun Baek, Seong Joon Oh, Seonghyeon Kim, Hwalsuk Lee.
Workshop on Text and Documents in the Deep Learning Era, CVPR WTDDLE 2020.
[Paper]

What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis
Jeonghun Baek, Geewook Kim, Junyeop Lee, Sungrae Park, Dongyoon Han, Sangdoo Yun, Seong Joon Oh, and Hwalsuk Lee.
ICCV 2019 oral (acceptance rate 4.3%, 187/4303).
[Paper] [Code].

International conference (refereed)

Descriptive, generative, and hybrid approaches for neural connectivity inference from neural activity data
Jeonghun Baek, Shigeyuki Oba, Junichiro Yoshimoto, Kenji Doya, and Shin Ishii.
Computational Neuroscience Meeting (CNS), Jul. 2016.

Computational Complexity Reduction for Functional Connectivity Estimation in Large Scale Neural Network
Jeonghun Baek, Shigeyuki Oba, Junichiro Yoshimoto, Kenji Doya, and Shin Ishii.
International Conference on Neural Information Processing (ICONIP), Nov. 2015.
[Paper]

A Situation-Aware Action Selection based on Individual’s Preference using Emotion Estimation
Kazumi Kumagai, Jeonghun Baek, and Ikuo Mizuuchi.
IEEE International Conference on Robotics and Biomimetics (ROBIO), Dec. 2014.
[Paper]

Considerate Behavior of Robots based on Individual’s Preference
Jeonghun Baek, and Ikuo Mizuuchi.
International Conference on Intelligent Autonomous System (IAS), Jul. 2014.
[Paper]

MISC.

ICDAR competition

• 1st place of ArT 2019 (2019.05.01) Chinese curved text recognition.
• 3rd place of ReCTS 2019 (2019.05.01) Chinese signboard text recognition.
• 3rd place of MLT 2019 (2019.06.03) Multi-lingual scene text detection and recognition.
• 3rd place of COCO-text 2017 (2019.02.25) Text recognition of COCO dataset, 1st if sorted by case sensitive.
• 1st place of Focused Scene Text 2013-2015 (2019.02.18) Typical scene text recognition.

-Last updated: Jul 8, 2020-