Me

PhD about Computer Vision and Machine Learning

I like the word 'mare', meaning sea in Latin.
('feelmare' is newly created my own words, because 'mare' word is always in use.)
I like computer games and programming since I was little. So I might be doing computer related job now.

I got a Doctor of Engineering degree in 2012 from pusan university in Korea.

My major is Computer Vision.
I am interested in 'Tracking', 'Machine learning', '3D reconstruction', and 'Vision odometry'.

I would like to always challenge the difficult issues in computer vision. And I can go anywhere in the world.
If you want to contact me, please email to me.

My email is feelmare@gmail.com, feelmare@daum.net.
facebook : https://www.facebook.com/studymarearts


Thank you!

2015.02.10





*Projects
Auto image categorizing on the PTX preset tour camera(2014.09~2014.11)
- The videos are stored as preset touring is very inconvenient when you want to see the video only particular scene.
- This algorithm can do categorize and distinguish the scene without any external information about the scene.
- see the more results on youtubehttps://www.youtube.com/watch?v=FP0zFOTI8t4
https://www.youtube.com/watch?v=KLbsjdiDMQc



Video Stabilizing(2014.07~2014.09)
- The project is to compensate unshakable video for shaky camera by calculating the amount of movement of the feature point in the image.
- The basic method is like that http://feelmare.blogspot.kr/2014/04/video-stabilizing-example-source-code.html, but more complex algorithm is required for normal environment.
- This technique is consisted of differential image, background image extraction(without moving object), feature extraction, matching(homography, ransac), affine transformation(movement, rotation), non-linear optimization, kalman filter...
- This subject is not easy problem, this project need more improvement.
- Please results refer to youtube.
https://www.youtube.com/watch?v=T-9G318vGNE
https://www.youtube.com/watch?v=vyrkas0s-Vw
https://www.youtube.com/watch?v=faOyels21lA




Torso and face detection in video(2014.04~2014.06)
- Search the upper body in the video, upper body means frontal and back.
- If face existed in the detected upper body, represent face in the result.
- Torso is learned by HOG+SVM and face is learned by Adaboost algorithm
- The rsult of torso+face is represented left side and torso only detection result is represented right side in the output video.
- see the result video in youtube




Real-time stitching multi-videos to one screen(2013.12~2014.03)
-The solution shows panorama image from multi-images. The panorama images is processing by real-time stitching algorithm.
- Each cameras has a limited field of view, but the solution can be monitoring large areas from merged into a panorama image.
-The solution can be applied efficiently and easy in Military Region, tourist attractions, intersections, ports
- See more detail in here.
http://feelmare.blogspot.kr/2015/02/real-time-stitching-multi-video-to-one.html




Real-time yard trailer identification by detection of vehicle ID number(2013.09~2013.11)
- Y/T(Yard Trailer) number identification solution using image processing
- The solution using camera is easy to installation and maintance compare to the RFID. And It is more free from distance constraint.
- The project used machine learning methods, SVM(Support Vector Machine), MLP(Multi Layers Perceptron), to recognize the number ID
- GPU used for a high-speed image processing.
- See more detail in here.
http://feelmare.blogspot.kr/2015/01/real-time-yard-trailer-identification.html





Traffic light detection and color recognition (2012.10~2013.01)
- I performed this project when I worked on GITC.
- The algorithm is to classify traffic light shape or not by HMM(hidden markov model) in candidate blob that is extracted by pre-trained color model.
- Demo video is here http://www.youtube.com/watch?v=OxCs1mCrLT8




AUV(Autonomous Underwater Vehicle) Project(2010.04~2012.07) - with POSTI (Postech Ocean Science & Technology Institute)
- Hardware Platform R&D, Software Architecture R&D(My Job), Autonomous(My Job)
- see the video on the youtube -> http://www.youtube.com/watch?v=AMxlSBvagy8




Unmanned Vehicle Project(2009.11~2010.10) -  Hyundai & Kia 10th Future Vehicle Contest(Link)
- Goal : To drive 4km course with unmanned vehicle and to accomplish various missions(The contest is similar to DARPA unmanned vehicle contest)
- Hardware Platform R&D, Software Architecture R&D(My Job), Autonomous(My Job): Obstacle Avoidance(Sensing, Localization, Mapping, Path planning)

 I introduced this project simply on my blog. -> Click



Humanoid Robot R&D(2007.3~2009.2) - with KIST(Korea Institute of Science and Technology)
*1st year :Real-Time Dynamic Path Panning
-2007.03.01~2008.02.29


*2st year :Robot and Interested Object Pose Estimation
-2008.03.01~2009.02.28

*3st year : Object pose estimation of Ellipse shape type
-2009.03.01~2010.02.28



Embeded camera R&D for Real-Time fire detection surveillance in tunnel environment(2008.7~2009.6)
- High Computing Power Embeded Camera R&D, Fire & Smoke Detection(My Job)
Introduced Page Link
 



SK+TU (SK telecom + Tongmyong University) Project(2006.3~2008.3)
*1st year : Object Based Video Communication
- 2006.03.15~2007.03.14
- Goal : To improve compression efficiency using extracting interested object in mobile communication.
- Image Processing in PDA(My Job), Fast Face detection algorithm R&D using CAMShift + AdaBoost Algorithm(My Job)

 *2st year : Intelligent Image Management System for Digital Photo Albums
- 2007.03.01~2008.02.29
- Goal : Intelligent Photo Album management Software R&D
- Multi-View Face Detection(My Job)



Machine Vision System R&D for detect defective parts of Car(2005.3~2006.3) - with SaeDong Corporation, Korea Intellignece Machine Corporation
- Goal : Machine Vision Hardware & Software R&D for detect defective parts of car.
- Hardware Production, Software  R&D, Auto Calibration in poor lighting condition(My Job)


*Publication
Domestic Conference.
[1] T. Zhu, J. H. Kim, D. J. Kang, M. S. Kim, and S. Lee, Ju, “영상감시시스템에서 은닉마코프모델을 이용한 불검출 방법,” 한국정보처리학회 춘계학술발표대회, vol. 16, no. 1, 2009.[ PDF ]
[2] J. Y. Lee, J. H. Kim, and D. J. Kang, “능동형상모델과 피부색 검출을 통한 손바닥경계 형상의 추적,” 한국정보처리학회 춘계학술발표대회, vol. 1, no. 13, pp. 681-684, 2006.[ PDF ]
[3] J. H. Kim, T. J. Rho, J. Y. Kim, and D. J. Kang, “적응적 Boosting 과 CAMShift 알고리즘을 이용한 실시간 고속 얼굴 검출 및 추적 시스템의 개발,” 전자공학회 추계 학술발표대회, 2006.[ PDF ]
[4] J. H. Kim, M. S. Kim, and D. J. Kang, “불균등 조명에서 비접촉 계측을 위한 반자동 카메라 교정 방법,” 한국정보처리학회 춘계학술발표대회, vol. 1, no. 29, pp. 281-282, 2006.[ PDF ]
[5] J. H. Kim, D. G. Kim, T. J. Lho, M. S. Kim, S. H. Lee, and D. J. Kang, “불균등 조명에서 비접촉 계측을 위한 자동 카메라 교정 방법,” 전자공학회 추계 학술발표대회, 2006. [ PDF ]
[6] Y. J. Yoon, J. H. Kim, D. J. Kang, "무인자동차 PtoP 주행에서 회전반경을 고려한 경로 추적에 관한 연구," 2010. [ PDF ]
[7] J. H. Kim, D. J. Kang, "스테레오 카메라와 차량 자세를 이용한 무인자동차의 주행 가능 영역 검출," 대한임베디드 공학회, 추계학술대회, 2012. 11. 08 [ PDF ]

Domestic Journal.
[1] J. H. Kim, T. Zhu, J. Y. Kim, and D. J. Kang, “혼합 약한 분류기를 이용한 AdaBoost 알고리즘의 성능 개선 방법,” Journal of Institute of Control, Robotics and Systems, vol. 15, no. 5, pp. 457-464, 2009. [ PDF ]
[2] J. H. Kim, Z. Teng, and D. J. Kang, “자기 조직화 지도를 이용한 다중 평면영역 검출,” Journal of Institute of Control, Robotics and Systems, vol. 17, no. 1, pp. 22-30, Jan. 2011. [ PDF ]
[3] J. H. Kim, T. J. Lho, J. Y. Kim, Y. J. Hong, J. W. Kwon, and D. J. Kang, “AdaBoost 기반의 실시간 고속 얼굴검출 및 추적시스템의 개발,” Journal of Control, Automation and Systems Engineering, vol. 13, no. 11, pp. 1074-1081, 2007. [ PDF ]

International Conference.
[1] J. H. Kim, T. Zhu, D. J. Kang, and J. E. Ha, “Multiple Plane Detection Method from Range Data of Digital Imaging System for Moving Robot Applications,” International Conference on Advanced Computational Engineering and Experimenting, 2010. [ PDF ]
[2] J. H. Kim, J. H. Park, and D. J. Kang, “METHOD TO IMPROVE THE PERFORMANCE OF THE ADABOOST ALGORITHM USING GAUSSIAN PROBABILITY DISTRIBUTION,” 2008 International Conference on Control, Automation and Systems, pp. 1749-1752, 2008. [ PDF ]
[3] J. H. Kim, J. H. Park, and D. J. Kang, “Camera calibration method under poor lighting condition in factories,” 2008 International Conference on Control, Automation and Systems, pp. 2162-2166, Oct. 2008. [ PDF ]
[4] J. H. Kim, B. G. Kwon, J. Y. Kim, and D. J. Kang, “Method to improve the performance of the AdaBoost algorithm by combining weak classifiers,” Content-Based Multimedia, pp. 357-364, 2008. [ PDF ]
[5] J. H. Kim, I. M. Ahn, and D. J. Kang, “Fast Cascade Training in AdaBoost Using Random Selection of Positive Data,” International Conferene on Mechatronics and Information Technology, pp. 56-58, 2009. [ PDF ]
[6] Z. Teng, J. H. Kim, and D. J. Kang, “Ellipse detection using an improved randomized Hough transformation,” IS&T/SPIE Electronic Imaging, Vol. 7870, pp. 78700U-78700U-8, San Francisco, USA, January, 2011. [ PDF is not prepared ]
[7] Z.Teng, J. H. Kim, and D. J. Kang, “Real-time lane detection by using multiple cues,” International Conference on Control, Automation and Systems, pp. 2334 - 2337, Seoul, South Korea, 2010. [ PDF ]
[8] S. C. Yu, J. H. Kim, T. Zhu and D. J. Kang, “Underwater Object Recognition Using 3D
Image Sonar for AUVs,” International Conference on WaterSide security, Singapore, 2012. [ PDF is not prepared]


International Journal.
[1] Z. Teng, J. H. Kim, and D. J. Kang, “Fire detection based on hidden Markov models,” International Journal of Control, Automation and Systems, vol. 8, no. 4, pp. 822-830, Aug. 2010. [ PDF ]
[2] J.H. Kim, Z. Teng, and D.J. Kang, “Real-time multiple plane area detection using a self-organizing map,” Optical Engineering, the journal of the Society of Photo-optical Instrumentation Engineers / 0036-1860 ; 0091-3286 ; 1560-2303 ; / 2012. [ PDF ]
[3] Z. Teng, J.H. Kim, and D.J. Kang, “Ellipse detection: A simple and precise method based on Randomized Hough Transform,” Optical Engineering,  the journal of the Society of Photo-optical Instrumentation Engineers / 0036-1860 ; 0091-3286 ; 1560-2303 ; / 2012 [ PDF ]
[4] J.H. Kim, Y. Z. Yoon, S. C. Yu, and D.J. Kang, “Real-time 2D height mapping method for an unmmaned vehicle using stereo camera and laser sensor fusion,” International Journal of Control, Automation and Systems, Vol. 10, no. 4, pp.761-771, 2012. [ PDF (manuscript version)]



Patent.
[1] 김정현, 강동중, 등죽, 이주섭, and 김민성, “영상 감시 시스템에서 은닉 마르코프 모델을 이용한 불 검출방법(fire detecting method using hidden markov models in video surveillance and monitoring system),” 출원번호: 1020060078283 (2006.08.18.), 등록번호: 100837119 (2008.06.04)
[2] 김정현, 김민성, 강동중, and 이세호, “영상 계측을 위한 카메라 교정방법(a camera calibration method for measuring the image),” 출원번호: 1020060078283 (2006.08.18.), 등록번호 : 100837119 (2008.06.04.)
[3] 김정현, 석수영, 박용완, “차량 위치 기반 교통 신호 검출 시스템 및 이를 이용한 교통 신호 거물 방법,” 출원번호: 10-2012-0155638(2012.12.28)
[4] 김정현, 이강석, "스포츠 중계를 위한 3차원 미니맵 서비스 제공장치(Providing apparatus of three dimensions mini map service for sports broadcasting)," 출원번호:1020140026241(2014.03.05), 등록번호:1014681590000(2014.11.26)