CAM00090.jpg

Huen Oh

Computer Vision expert with 5+ years of professional experience in ADAS system of automotive fields, concentrated on detection and recognition of specific objects using Machine Learning technologies.

 - 6+ yrs of C/C++ experience developing Computer Vision and Machine Learning algorithms.

 - Strong understanding of algorithms and data structures, with analytic approaches.

 - Quick learner and self motivated problem solver.

 - Highly adaptable in any programming languages. (Experienced in Python, Java, HTML, MATLAB, etc.) .

Experience

Research engineer,  Jun 2012 - Present

@ Machine Learning Recognition Group in DAS Sensor Engineering Team

< Working Projects >

  • Traffic Sign Recognition System (EU, USA): Ready to in product.
  • Vehicle Detection. : In production.

< Details >

  • Logic Development (C/C++)
    • Features: Modified Ternary, HOG, LBP, MCT, YUV Histogram
    • Training Techniques: Adaboost, SVM, Deep Learning
    • Responsible Modules: Decision Tree, Classification, Training, False positive rejection.
    • Logic simulation and Field test
    • Code Refactoring and Error fix written by other colleagues
  • Set up training process (C/C++, Python, MATLAB)
    • Set up and run a chain of training process from scratch.
      Patch Generation -> Feature Extraction -> Learning Process + Bootstrapping
    • Write, run and validate training sessions.
  • Tooling (Visual Studio 2013, MFC, C/C ++, Python, OpenCV)
    • PC version framework to run and test logic
    • Ground Truth Annotation Tool(TinyXml, CAN), Feature Extraction Management Tool, Automatic Classifier Test Tool, Tree Node Analysis Tool(Graphviz), Ground truth verification Tool (Python, Qt)
  • Deep Learnings
    • Following Up-to-date trends : Capsule-Net, YOLO v3, GAN
    • Code-level understanding : Hands on coding experience from scratch(Python), Darknet

mobis.png

Computer Vision SW Designer,  Jul 2011 - May 2012

Development of Vision Align module in Film Attach Machine for LG smartphone.

techwin.jpg

Research Assistant,  Jan 2011 - May 2011

@ Computational Breast Image Group in Radiology Department

  • Breast MRI Data Acquisition from PACS(Picture Archiving and Communication System).
  • Fields modifications of DICOM (Digital Imaging and Communications in Medicine) to fit into a different archive system. (e.g. ClearCanvas).

upenn.png

Intern,  Jun 2010 - Aug 2010

Mechanical Design and Development to maximize the capabilities of the VICON camera system through an in-depth understanding and ultimate write-up of the accuracy both inside and outside the building, optimal camera configuration, dead spots, etc..

motile_robotics.gif




Education

UNIVERSITY OF PENNSYLVANIA, Philadelphia, PA

  • Master of Science in Robotics, May 2011
    • Simultaneous Localization and Mapping with Kinect Sensor (SLAM) (MATLAB, C++, 2011) 
      Created map and localized robot simultaneously from the sensor data. (LIDAR, accelerometer and gyro) Unscented Kalman Filter for IMU orientation and Particle Filter for SLAM are used. 3D mapped with Kinect depth image over 2D map.
    • Automatic Logo Replacement (MATLAB, 2010)
      Replaced logo automatically using Shape Context, HoG, Thin Plate Spline, Homography and Adaptive RANSAC.
    • Jigsaw Puzzle Solver (MATLAB, 2010)
      Solved Jigsaw puzzle problems with an image of scattered pieces and an original image using Edge detection, Segmentation and SIFT. Found the relationships, proper locations and orientations of pieces and constructed a solved picture.
    • Developing age and gender prediction algorithms (MATLAB, 2009)
      Trained machine and made prediction of age and gender from blog postings (text) and faces (images). Applied and analyzed SVM, PCA, Kernel Logistic Regression, Gaussian Naive Bayes, AdaBoost with weighted Naive Bayes to the data. Used Cross Validation to check the estimated result. 2 members in one team. Got 1st place in the Machine Learning class.
    • Mobile robot slalom using vision recognition (MATLAB, 2009)
      Controlled a 3-wheels mobile robot to pass through the gates as fast as possible using vision detection with a web camera on top of the robot. Used HSV and RGB color to distinguish objects and Geometric approach to calculate distance and angle of the objects.

KYUNG HEE UNIVERSITY, Seoul, The Republic of Korea

  • Bachelor of Engineering in Computer / Electronic Engineering (Double Major), February 2009 
    • Thesis - Elimination of Video Redundancy by Axis Rearrangement, December 2007
  • Research Assistant, Database and Knowledge Engineering Lab, Jun/2006 ~ Jun/2007
    • Conducted research in project: Development of Smart Multi-party Interactive Collaboration System in Ubiquitous Community Environments
    • Huen Oh, Jin-Seung Kim, Young-Koo Lee, “Effective Method for Data Sharing on Access Grid Using Keyword-Based Full-text Retrieval,” Korea Information Processing Society (KIPS), May 2006.

Skills


Proficient in

C, C++, MFC, MATLAB, Git, Machine Learning, Computer VisionDeep Learning, Code Re-factoring, Debugging, CAN


Experienced in

Linux, Python, Qt, Java, HTML, Open Libraries(OpenCV, TynyXml, Graphviz, Tensorflow) 


Interested in

Deep Learning, Robotics, VR, HMI.

Powered by CakeResumePowered by CakeResume
Powered by CakeResumePowered by CakeResume