Kwak, Nojun_Affiliated Professor

RESEARCH INTERESTS

Machine learning algorithms and its applications to pattern recognition and computer vision.

I have been mainly working on statistical analysis of input variables for pattern recognition problems since my graduate studies. Especially, my main research interest was to develop feature selection and feature extraction algorithms for pattern recognition problems such as face recognition and image registration. It requires thorough knowledge on various of study such as pattern recognition, machine learning, artficial neural networks, data mining and image processing.

Currently, I am interested in the theoretical aspect of kernel methods in machine learning algorithms. Also, I am interested in detecting and tracking objects in real world environments by using computer vision algorithms and their applications for autonomous vehicles and robot applications. In 2015, I have participated DARPA robotics challenge as a co-leader of the Team-SNU.

Nowadays, deep neural networks are the main area of interest in my group. We deal with theoretical aspects as well as the applications of deep neural networks for various problems.

Education

Seoul National University Mar. 1999 – Feb. 2003
Ph.D. in Electrical and Computer Engineering
Supervisor: Prof. Chong-Ho Choi
Thesis: Feature selection and extraction based on mutual information for classication

Seoul National University   Mar. 1997 – Feb. 1999
M.S. in Electrical Engineering
Supervisor: Prof. Chong-Ho Choi
Thesis: Input feature selection for neural networks using mutual information and Taguchi method

Seoul National University  Mar. 1993 – Feb. 1997
B.S. in Electrical Engineering

University of Illinois at Urbana-Champaign   Jul. 2001 – Jun. 2002
Visiting Graduate Student in Electrical and Computer Engineering
Supervisor: Prof. Narandra Ahuja

EXPERIENCE

Seoul National University   Sep. 2013 – Present
Associate Professor
Department of Transdisciplinary Studies,
Graduate School of Convergence Science and Technology

Ajou University    Mar. 2007 – Aug. 2013
Assistant/Associate Professor
Department of Electrical Engineering

Seoul National University     Sep. 2006 – Feb. 2007
BK Assistant Professor
BK (Brain Korea) Information Technology Research Group

Samsung Electronics      Mar. 2003 – Aug. 2006
Senior Engineer
Home-network Lab.
Standardization Lab.
Worked as a Samsung Delegate for 3GPP RAN3

BOOK

H.-C. Shin, H.-B. Kang, N. Kwak, H.-I. Kim, S.-B. Cho, Imaging and vision systems for intelligent vehicle safety (Korean edition), HongReung Science Publisher, ISBN 9788997570492, 2012.

Publication

1. Nojun Kwak, “Implementing kernel methods incrementally by incremental nonlinear projection trick”, IEEE Transactions on Cybernetics, accepted for publication.
2. Jiyong Oh and Nojun Kwak, “Generalized mean for robust principal component analysis”, Pattern Recognition, vol. 54, pp. 116-127, June 2016.
3. Eunwoo Kim, Minsik Lee, Chong-Ho Choi, Nojun Kwak, and Songhwai Oh, ” Ecient L1-Norm- Based Low-Rank Matrix Approximations for Large-Scale Problems Using Alternating Rectied Gradient Method”, IEEE Transactions on Neural Networks and Learning Systems, vol. 26, no. 2, pp. 237 – 251, Feb. 2015.
4. Nojun Kwak Principal component analysis by Lp-norm maximization, IEEE Transactions on Cybernetics, vol.44, no. 5, pp. 594 – 609, May 2014.
5. Nojun Kwak, ” Nonlinear Projection Trick in Kernel Methods: An Alternative to the Kernel Trick”, IEEE Transactions on Neural Networks and Learning Systems, vol. 24, no. 12, pp. 2113 – 2119, Dec. 2013.
6. Jiyong Oh, Nojun Kwak, Minsik Lee and Chong-Ho Choi, Generalized mean for feature extraction in one-class classi cation problems, Pattern Recognition, vol. 46, no. 12, pp. 3328 – 3340, Dec. 2013.
7. Jae Hyun Oh and Nojun Kwak, Generalization of linear discriminant analysis using Lp-norm, Pattern Recognition Letters, vol. 34, no. 6, pp. 679 – 685, April 2013.
8. Sang-Il Choi, Chong-Ho Choi, Gu-Min Jeong and Nojun Kwak, Pixel selection based on discrim- inant features with application to face recognition, Pattern Recognition Letters, vol. 33. no. 7, pp. 1083 – 1092, July 2012.
9. Jae Hyun Oh and Nojun Kwak, Recognition of a driver’s gaze for vehicle headlamp control, IEEE Transactions on Vehicular Technology, vol. 61, no. 5, pp. 2008 – 2017, June 2012.
10. Nojun Kwak, Kernel discriminant analysis for regression problems, Pattern Recognition, vol. 45, no. 5, pp. 2019 – 2031, May 2012.
11. Jungtai Kim, Hyun Jong Yang and Nojun Kwak, Low-angle tracking of two objects in three- dimensional beamspace domain, IET Radar, Sonar & Navigation, vol. 6, no. 1, pp. 9-20, Jan. 2012.
12. Eun-Chan Park, Nojun Kwak, Suk Kyu Lee, Jong-Kook Kim and Hwangnam Kim, Provision- ing QoS for WiFi-enabled Portable Devices in Home Networks, KSII Trans. on Internet and Information Systems, vol. 5, no. 4, pp. 720 – 740, April, 2011.
13. Sang-Il Choi, Chong-Ho Choi and Nojun Kwak, Face recognition based on 2D images under illumination and pose variations, Pattern Recognition Letters, vol. 32, pp. 561 – 571, March 2011.
14. Nojun Kwak and Jung-Won Lee Feature extraction based on subspace methods for regression problems, Neurocomputing, vol. 73, no. 10, pp. 1740 – 1751, June 2010.
15. Gu-Min Jeong, Hyun-Sik Ahn, Sang-Il Choi, Nojun Kwak, and Chanwoo Moon, ” Pattern Recogni- tion Using Feature Feedback: Application to Face Recognition”, International Journal of Control, Automation, and Systems, vol. 8, no. 1, pp. 141 148, Jan. 2010.
16. Nojun Kwak and Jiyong Oh, Feature extraction for one-class classi cation problem: Enhance- ments to biased discriminant analysis, Pattern Recognition, vol. 42, no. 1, pp. 17 – 26, Jan. 2009.
17. Nojun Kwak, Principal component analysis based on L1 norm maximization, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 30, no. 9, pp. 1672 – 1680, Sep. 2008.
18. Nojun Kwak, Chunghoon Kim and Hwangnam Kim, Dimensionality reduction based on ICA for regression problems, Neurocomputing, vol. 71, no. 13, pp. 2596 – 2603, Aug. 2008.
19. Nojun Kwak, Feature extraction for classi cation problems and its application to face recognition, Pattern Recognition, vol. 41, no. 5, pp. 1701 – 1717, May 2008.
20. Nojun Kwak, Feature extraction based on direct calculation of mutual information, International Journal of Pattern Recognition and Arti cial Intelligence, vol. 21, no. 7, pp. 1213 – 1231, Nov. 2007.
21. Chong-Ho Choi and Nojun Kwak, Robust control of robot manipulator by model-based distur- bance attenuation, IEEE/ASME Transactions on Mechatronics, vol. 8, no. 4, pp. 511 – 513, Nov./Dec. 2003.
22. Nojun Kwak and Chong-Ho Choi, Feature extraction based on ICA for binary classi cation problems, IEEE Transactions on Knowledge and Data Engineering, vol. 15, no. 6, pp. 1374 – 1388, Nov. 2003.
23. Nojun Kwak and Chong-Ho Choi, Input feature selection by mutual information based on Parzen window, IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 24, no. 12, pp. 1667-1671, Dec. 2002.
24. Nojun Kwak and Chong-Ho Choi, Input feature selection for classi cation problems, IEEE Trans- actions on Neural Networks, vol. 13, no. 1, pp. 143 – 159, Jan. 2002.

INTERNATIONAL
CONFERENCES

Sungheon Park, Jihye Hwang and Nojun Kwak, “3D Human Pose Estimation Using Convolu- tional Neural Networks with 2D Pose Information”, ECCV2016 Workshop (Geometry Meets Deep Learning), Accepted, 2016. (arXiv:1608.03075)
Sungheon Park and Nojun Kwak, “Analysis on the Dropout E ect in Convolutional Neural Net- works”, Proc. Asian Conf. on Computer Vision (ACCV2016), Accepted, 2016.
Myunggi Lee, Hyeogjin Lee, Jiyong Oh, Hak Jong Lee, Seung Hyup Kim and Nojun Kwak, “Unregistered Bosniak Classi cation with Multi-phase Convolutional Neural Networks”, Proc. Int’l Conf. on Neural Information Processing (ICONIP2016), Accepted, 2016.
Jihye Hwang, Younggwang Ji, Nojun Kwak and Eun Yi Kim, ” Outdoor context awareness device that enables mobile phone users to walk safely through urban intersections”, Proc. Int’l Conf. on Pattern Recognition Applications and Methods (ICPRAM’16), Rome, Italy, Feb. 2016.
John Yang, Hyeogjin Lee and Nojun Kwak, ” Kernel Parameter Selection by Gap Maximization between Intra and Inter{Class Samples”, Third International Conf. on Big Data and Smart Computing (BigComp’16), Hong Kong, China, Jan. 2016.
Sungheon Park, Myunggi Lee, Jihye Hwang, John Yang, Jieun Lee and Nojun Kwak, “Object Localization Using Image Classi cation Score” poster presentation at ICCV2015 Workshop (Im- ageNet and MS COCO Visual Recognition Challenges Joint Workshop), Santiago, Chile, Dec. 2015.
Sanghyun Kim, Mingon Kim, Jimin Lee, Soonwook Hwang, Joonbo Chae, Beomyeong Park, Hyunbum Cho, Jaehoon Sim, Jaesug Jung, Hosang Lee, Seho Shin, Minsung Kim, Nojun Kwak, Youngjin Lee, Sangkuk Lee, Myunggi Lee, Sang Yup Yi, Kyong{Sok K.C. Chang, Jaeheung Park, “Approach of Team SNU to the DARPA Robotics Challenge Finals”, 2015 IEEE{RAS Int. Conf. on Humanoid Robots (Humanoid’15), Seoul, Korea, November 2015.
Sungheon Park and Nojun Kwak, ” Illumination robust opticalw estimation by illumination-chromaticity decoupling”, IEEE International Conf. on Image Processing (ICIP’15), Quebec City, Canada, September 2015.
Hyeogjin Lee and Nojun Kwak, ” Character recognition for the machine reader zone of electronic identity cards”, IEEE International Conf. on Image Processing (ICIP’15), Quebec City, Canada, September 2015.
Minsik Lee, Jieun Lee, Hyeogjin Lee, and Nojun Kwak, ” Membership representation for detecting
block{diagonal structure in low{rank or sparse subspace clustering” , IEEE Conf. Computer Vision and Pattern Recognition (CVPR’15), Boston, MA, June 2015.
Sungheon Park and Nojun Kwak, ” Cultural event recognition by subregion classi cation with con- volutional neural network”, Workshop in IEEE Conf. Computer Vision and Pattern Recognition (CVPR’15), Boston, MA, June 2015.
Sungheon Park, Myunggi Lee and Nojun Kwak, ” Polyp detection in colonoscopy videos using
deeply{learned hierarchical features”, Presented at ISBI 2015 Challenge on Automatic Polyp De- tection in Colonoscopy Videos, New York, NY, April 2015.
Kanishka Tyagi, Nojun Kwak and Michael Manry,Optimal conjugate gradient algorithm for generalization of linear discriminant analysis based on L1 norm , Proc. Int’l Conf. on Pattern Recognition Applications and Methods (ICPRAM’14), pp.207 – 212, Angers, France, March2014.
Seaung-Lok, Ham and Nojun Kwak, Boosted{PCA for binary classi cation problems, Proc. IEEE Int. Symposium on Circuits and Systems (ISCAS’12), Seoul, Korea, May 2012.
Ji-Eun Lee and Nojun Kwak, Detection and recovery of occluded face images based on correlation between pixels, Proc. 1st International Conference on Pattern Recognition Applications and Methods (ICPRAM’12), vol. 2, pp. 568 – 571, Vilamoura, Algarve, Portugal, February, 2012.
Jae Hyun Oh and Nojun Kwak, Robust face recognition under a polar coordinate system, Proc. the 2011 International Conference on Image Processing, Computer Vision, & Pattern Recognition (IPCV11), Las Vegas, NV, July 2011.
Jae Hyun Oh and Nojun Kwak, Robust face recognition in a polar coordinate system using LDAr, Proc. the 2nd International Conf. on Internet (ICONI10), Mactan Island, the Philippines, Dec. 2010.
Jiyong Oh, Chong-Ho Choi and Nojun Kwak, MAP Classi er with BDA features, Proc. the 6th International Conference on Fuzzy Systems and Knowledge Discovery (FSKD’09), pp.227- 231, Tianjin, China, Aug. 2009.
Nojun Kwak, Sang{Il Choi, and Chong-Ho Choi, Feature extraction for regression problems and an example application for pose estimation of a face, Proc. Intl Conf. on Image Analysis and Recognition (ICIAR08) (Also presented in Lecture Notes in Computer Science, vol. 5112), pp. 435 { 444, Povoa de Varzim, Portugal, July 2008.
Nojun Kwak, Feature extraction with weighted samples based on independent component analysis, Proc. Intl Conf. on Arti cial Neural Networks (ICANN06) (Also presented in Lecture Notes in Computer Science, vol. 4132), pp. 340 – 349, Athens, Greece, Sep. 2006.
Nojun Kwak and Chunghoon Kim, Dimensionality reduction based on ICA for regression prob- lems, Proc. Intl Conf. on Arti cial Neural Networks (ICANN06) (Also presented in Lecture Notes in Computer Science, vol. 4131), pp. 1 – 10, Athens, Greece, Sep. 2006.
Nojun Kwak and Chong-Ho Choi, Feature extraction using mutual information based on Parzen window, Proc. Intl Conf on Arti cial Neural Network and Neural Information Processing (ICANN/ ICONIP03), pp. 278 – 281, Istanbul, Turkey, June 2003.
Nojun Kwak, Chong-Ho Choi, and Narendra Ahuja, Face recognition using feature extraction based on independent component analysis, Proc. Intl Conf. on Image Processing (ICIP02), vol. 2, pp. 337 – 340, Rochester, New York, Sep. 2002.
Nojun Kwak, Chong-Ho Choi, A new method of feature extraction and its stability, Proc. Intl Conf. on Arti cial Neural Networks (ICANN02) (Also presented in Lecture Notes in Computer Science, vol. 2415) pp. 480 – 485, Madrid, Spain, Aug. 2002.
Nojun Kwak, Chong-Ho Choi, and Jin{Young Choi, Feature Extraction using ICA, Proc. Intl Conf. on Arti cial Neural Networks (ICANN01) (Also presented in Lecture Notes in Computer Science, vol. 2130), pp. 568 – 573, Vienna, Austria, Aug. 2001.
Chong-Ho Choi and Nojun Kwak, Disturbance attenuation in robot control, Proc. Intl Conf. on Robotics and Automation (ICRA01), vol. 3, pp. 2560 – 2565, Seoul, Korea, May 2001.
Nojun Kwak and Chong-Ho Choi, Improved mutual information feature selector for neural net- works in supervised learning, Proc. Int’l Joint Conf. on Neural Networks (IJCNN99), vol. 2, pp.
1313 – 1318, Washington D.C., July 1999.

INVITED TALKS

Maximizing L1 norm in data analysis, in The First International Symposium on System Informatics and Engineering, Qingdao, China, 2011.
L1 norm optimization in subspace learning methods, in AI-2011 Thirty-first SGAI International Conference on Arti cial Intelligence, Cambridge, UK, 2011

OTHERS

IEEE Senior Member
8 Korean Journal papers
23 Korean Domestic Conference papers
Around 30+ registered patents.