Computer Vision Machine Learning Lab

Junseok Kwon, Assistant Professor
School of Computer Science and Engineering, Chung-Ang University


221 Hueksuk-Dong, Dongjak-gu, Seoul
156-756, Korea
Tel: +82-2-820-5914
Office: 310-735
E-mail: jskwon@cau.ac.kr

Google scholar




News

- I am serving as a Area Chair for ACCV 2018
- I am serving as a Finance Chair for ACM Multimedia 2018
- Outstanding Reviewer Award, BMVC 2017
- My paper entitled "visual tracking decomposition" published in CVPR 2010 has been cited more than 1000 times.
- My first paper with my student is accepted at BMVC 2017 -- congrats guisik!
- I am looking for Undergraduate Interns and Graduate Students.


Major Interests

- Deep learning
- Objectness
- Object tracking
- Dehazing
- Visual surveillance
- Event detection and summarization
- Monte Carlo Sampling method and its variants


Members

- Graduate students
Guisik Kim (2017.09~, PhD Candidate)
Sung Woo Park (2018.03~, Master)

- Undergraduate Interns
, Ji Young Park, Kevin Jin, Ga Young Kim, Jaehyun Lim, Minseok Song, ӹα, Sumngmin Cho, Dokyeong Kwon, Jinhee Park, Bo Won Choi, Suhyeon Ha, ȫ, (Undergraduate, CAU)

- Collaborator
Janghoon Choi, Sungyong Baik (PhD candidate, SNU)
Radu Timofte (Post-doctoral researcher, ETHZ)

- Alumni
Chan-Ju Park (2017, Qualcomm), Rae Hyuk Jung (2018, KAIST)


Professional Activities

- Editor: Member of the Editorial Committee, ETRI Journal (2017.08~)

- Area Chair: ACCV (2018)

- FInance Chair: ACM Multimedia (2018)

- Reviewer: TPAMI, TMM, TIP, TCSVT, IJCV, CVIU, PR, TII, SPL, IVC, JEI, JVCI, JIVP, SPI, KNOSYS, NEUCOM, CMPB, IJBI, DSP, JR, CVPR (2012~2018), ICCV(2011, 2013, 2015, 2017), ECCV (2012, 2014, 2016), ACCV (2010, 2012, 2014, 2016), BMVC (2017)


Recent & Selected Publications [Click here for a full list]

- Uncertainty Calibrated Markov Chain Monte Carlo Sampler for Visual Tracking based on Multi-Shape Posterior
Junseok Kwon, Journal of Mathematical Imaging and Vision (JMIV), accepted, 2018
[pdf]

- Visual Tracking Based on Edge Field with Object Proposal Association
Junseok Kwon, Hansung Lee, Image and Vision Computing (IVC), vol. 69, January, page 22-32. 2018
[pdf]

- Robust Pixel-wise Dehazing Algorithm based on Advanced Haze-Relevant Features
Guisik Kim, Junseok. Kwon, The 28th British Machine Vision Conference (BMCV), London, UK. 2017.
[pdf]

- Leveraging Observation Uncertainty for Robust Visual Tracking
Junseok Kwon, Radu Timofte, and Luc Van Gool, Computer Vision and Image Understanding (CVIU), vol. 158, page 62-71. 2017
[pdf]

- Adaptive Visual Tracking with Minimum Uncertainty Gap Estimation
Junseok Kwon, and Kyoung Mu Lee, IEEE Transaction on Pattern Analysis and Machine Intelligence (TPAMI), vol. 39, no. 1, page 18-31. 2017
[pdf]

- Tracking by Switching State Space Models
Junseok Kwon, Ralf Dragon, and Luc Van Gool, Computer Vision and Image Understanding (CVIU), vol. 153, page 29 - 36. 2016
[pdf]

- PICASO: PIxel Correspondences And SOft Match Selection for Real-time Tracking
Radu Timofte, Junseok Kwon, and Luc Van Gool, Computer Vision and Image Understanding (CVIU), vol. 153, page 162 - 153. 2016
[pdf]

- A Unified Framework for Event Summarization and Rare Event Detection from Multiple Views
Junseok Kwon, and Kyoung Mu Lee, IEEE Transaction on Pattern Analysis and Machine Intelligence (TPAMI), vol. 37, no. 9, page 1737-1750. 2015
[pdf]

- Tracking by Sampling and Integrating Multiple Trackers
Junseok Kwon, and Kyoung Mu Lee, IEEE Transaction on Pattern Analysis and Machine Intelligence (TPAMI), vol. 36, no. 7, page 1428-1441. 2014
[pdf]

- Appearances can be deceiving: Learning visual tracking from few trajectory annotations
Santiago Manen, Junseok Kwon, Matthieu Guillaumin, and Luc Van Gool, The 13th European Conference on Computer Vision (ECCV), Zurich, Switzerland. 2014.
[pdf] [project page]

- Robust Visual Tracking with Double Bounding Box Model
Junseok Kwon, Junha Roh, Kyoung Mu Lee, and Luc Van Gool, The 13th European Conference on Computer Vision (ECCV), Zurich, Switzerland. 2014.
[pdf]

- Interval Tracker: Tracking by Interval Analysis
Junseok Kwon, and Kyoung Mu Lee, The 27th IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Ohio, USA. 2014.
[pdf]

- Highly Non-Rigid Object Tracking via Patch-based Dynamic Appearance Modeling
Junseok Kwon, and Kyoung Mu Lee, IEEE Transaction on Pattern Analysis and Machine Intelligence (TPAMI), vol. 35, no. 10, page 2427-2441. 2013
[pdf]

- Wang-Landau Monte Carlo-based Tracking Methods for Abrupt Motions
Junseok Kwon, and Kyoung Mu Lee, IEEE Transaction on Pattern Analysis and Machine Intelligence (TPAMI), vol. 35, no. 4, page 1011-1024. 2013
[pdf] [project page]

- Minimum Uncertainty Gap for Robust Visual Tracking
Junseok Kwon, and Kyoung Mu Lee, The 26th IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Portland, USA. 2013.
[pdf]


Project

- The technology development for event recognition/relational reasoning and learning knowledge based system for video understanding
Funded by IITP, Korea (July 2017 - December 2021)
- Long-term visual tracking system based on deep learning
Funded by NRF, Korea (March 2017 - February 2020)
- Visual tracking with context and fusion
Funded by Chung-Ang University (March 2016 - February 2018)


Teaching

- Calculus (49149-01, Spring 2018)
- Machine vision applications (297131-01, Spring 2017)
- Engineering Mathematics (02331-01, Spring 2017)
- Probability and statistic (30195-01, 30195-02, Fall 2016, 30195-01, 30195-02, Fall 2017)
- Linear Algebra (11185-01, 11185-02, Spring 2016, 11185-01, 50153-01, Spring 2017, 11185-01, Spring 2018)


Educations

2013.6 ~ 2014.12: Post-doctoral researcher in D-ITET, ETH Zurich, Switzerland.
2013.3 ~ 2013.6: Post-doctoral researcher in ASRI, SNU, Seoul, Korea.
2008.3 ~ 2013.2: Doctor's course in EECS, SNU, Seoul, Korea.
2006.3 ~ 2008.2 : Master's course in EECS, SNU, Seoul, Korea.


Awards

2017.09 28th BMVC Outstanding Reviewers
2014.11 12th ACCV Best Reviewers
2012.11 11th ACCV Best Reviewers
2012.6 25th CVPR Doctoral Consortium
2012.2 18th Samsung Humantech Thesis Prize : Bronze Prize
2011.2 23th Image Processing and Image Understanding : Best Poster
2011.2 17th Samsung Humantech Thesis Prize : Gold Prize
2010.2 16th Samsung Humantech Thesis Prize : Honor Prize


Patents

Junseok Kwon and Kyoung Mu Lee. KOR Patent 1011077360000, Method for tracking object on visual (January, 2012)