Home: 1724 Victorian Court, Alpharetta, GA
Temporary home: 7421 E. Mckinley St. Scottsdale, AZ
Cell: (773)620-6559, Ping_dong99@yahoo.com
Looking for a computer vision and machine learning related position.
SUMMARY OF EXPERTISE
• Computer Vision and Machine Learning
◦ Object classification based on bag of words
◦ Machine leaning algorithms: SVM, Neural Networks (MLP), k-NN, decision tree, and Adaboost.
◦ Optimization algorithms: linear and non-linear optimization, dynamic programming, genetic algorithm.
◦ In-depth knowledge of random processes and statistical model parameter estimation: generative and discriminative model, Gaussian mixture model, EM algorithm, MAP, etc.
◦ Libraries: OpenCV, and LibSVM.
◦ Knowledge of deep neural networks.
• Digital Signal Processing
◦ Image, audio and video signal processing: filtering, restoration, estimation, prediction, detection, voice activity detection, and motion estimation.
◦ Speech signal processing: LPC based vocoders, speech recognition.
◦ In-depth knowledge and experience on media codecs: MPEG2/4, H.264, AAC.
◦ Adaptive filtering: LMS, RLS, GAL.
◦ Echo cancellation and noise suppression.
◦ Fixed-point and floating-point DSP algorithms: TMS320C64xx.
◦ Medical image reconstruction: SPECT, PET, filtered back projection, algebraic reconstruction.
• Programming Skills
◦ Languages: C, C++, assembly, Perl, Python, B/Cshell.
◦ Algorithm prototyping and simulation with Matlab.
◦ Unix/Linux system programming: Linux kernel, IPC, Socket.
◦ Parallel programming: openMP, Pthread.
◦ UML design with Enterprise Architect and code generation with MDworkbench.
• Software Development Process
◦ Source code revision control: Rational Team Concert (RTC), CVS, Perforce, MS Visual SourceSafe, ClearCase, GIT, and Subversion (SVN).
◦ Defect tracking software: RTC, ClearQuest, Remedy, Bugzilla.
◦ Code review: Code Collaborator.
◦ Static source code analysis: Coverity desktop plugin.
◦ Software process improvement approach: SCRUM, CMMI.
◦ Software quality assurance (SQA): functional test, load test, regression test, script based testing automation.
• Integrated Development Environment (IDE): MS Visual Studio, QNX Momentics, Code Composer Studio, Keil MDK-ARM tools.
PROFESSIONAL WORKING EXPERIENCE
Advanced Analytics Engineer (contractor), American Traffic Solution, Mesa, Arizona (Oct, 2015 –now)
Work on vehicle make and model classification and color recognition for automatic traffic data processing
• Worked on vehicle color recognition project.
• Worked on vehicle make and model classification based on bag of words (BoW) model. Dense SIFT feature descriptors, kmeans clustering algorithm, LLC coding and pooling, and SVM are used in the project. The algorithm can achieve 90% accuracy even on bad quality image data for large number of vehicle model classes (several hundred model year classes).
• Researched Eigen-face and fisher face methods for vehicle classification unsuccessfully before finding BoW methods.
• Researched on license plate locating based on morphological image processing and license number recognition based on OCR.
• Researched preprocessing of images, such as contrast enhancement, homography transform based view correction.
• Research is done on Matlab platform, and final software is developed on C++ and OpenCV library.
• Did performance tuning of the software by using OpenMP parallel programming.
• Technologies used: Lab color space, SIFT, SURF, SVM, k-NN, kmeans, C++, OpenCV, Matlab, OpenMP, image processing, OCR (Tesseract-ocr), MS visual studio, and XML data serialization.
Software engineer (contractor), Panasonic Automotive Systems, Peachtree City, Georgia (Feb, 2015-Aug, 2015)
Work as software developer and domain integrator for navigation application for a major car manufacturer
• New feature implementation and bug fix for navigation service on QNX RTOS.
• Software integration and promotion based on RTC team collaboration tool.
• Work as technical coordinator with third-party technology provider for head unit localization for Chinese market.
• Technologies used: multithreading based on Pthread; POSIX IPC based on message queue; QNX based embedded application development; GPS signal and Gyro, accelerometer, and wheel pulse based dead-reckoning, map matching for navigation; Rational Team Concert based software lifecycle management; C code based legacy framework combined with C++ based services, UML based interface design.
Principal Engineer, DeseamLab Inc., Beijing, China (2006-2013)
Lead various projects in signal processing and networking products
• Led video stabilization project. The project aims to remove annoying movement caused by the shaking camera. Researched algorithms for motion parameter estimation and designed parameter smoothing algorithms using LMS in time domain.
• Implemented some object detection algorithms for video monitoring, including car and human detection, fire detection.
• Researched music feature extraction, which is used in music classification project. Features such as tempo, key, pitch, chord, spectral, and vocal/non-vocal information are extracted and used in a clustering algorithm.
• Researched the acoustic echo cancellation problem. Implemented and tested some algorithms under different scenarios.
• Technologies used: Matlab based signal processing algorithm prototyping; C, C++ and Python coding; Audio and image signal processing; Signal classification based on clustering algorithms.
Senior Research Associate and PhD candidate, Imaging Research Center, Illinois Institute of Technology, Chicago, Illinois (2002-2006)
Finished my dissertation in digital image processing area, and worked in the lab as postdoc fellow for one year
• Conducted research on SPECT and PET image reconstruction algorithms (both analytic and iterative). Implemented cardiac motion estimation algorithms based on deformable mesh model of the images.
• Conducted research in digital image watermarking, focusing on geometric robust watermarking.
• Explored and experimented many algorithms in related areas, such as machine learning and computer vision.
• Technologies used: Signal detection, estimation, filtering, prediction; Image registration, segmentation, edge and feature detection; computer vision; Machine learning algorithms; Signal modulation, demodulation; Medical image reconstruction; Coding in Matlab, C, C++, Python and Perl.
Lead Embedded Software Engineer, RF subscriber unit, Motient Corp., Lincolnshire, Illinois (2001-2002)
Radio Network Controller (RNC) Development – pSOS based embedded system development
• Enhanced and maintained the company’s national wireless data network, which employing the DataTAC protocol deployed in the 800MHz band. Responsible for data traffic, capacity and latency analysis, bug fix, adding new features and software release.
• Enhanced the loadbox-testing tool to simulate more scenarios in traffic patterns.
• Developed many Tcl/Expect scripts to automate testing.
• Technologies used: C and pSOS RTOS based embedded application development; Wireless data network protocols; other networking protocols, such as TCP/UDP/IP, X25, Ethernet, socket programming.
Software Engineer: Voice Solution Department, 3Com, Rolling Meadows, Illinois (2000-2001)
IP phone product development – develop scripts for automated test and data processing
• Developed a signaling server load-testing tool using Perl. It is used for SIP protocol based call generation, assuming Poison distribution of the incoming calls. Performance of the proxy servers is measured by the distribution of the response time under specified load condition.
• Developed tools to do post-processing of testing data using Perl and gnuplot. The tools collect the load testing data by processing the log files, calculate the statistics (such as mean, max, min and standard deviation of call setup and teardown time, failure rate of call setup and teardown), and post the statistics and plots of the call distribution to an internal website.
• Developed scripts to automate the regression test.
• Conducted voice quality evaluation by both subjective test (MOS) and objective evaluation of speech distortion by comparing with the reference signal.
• Technologies used: SIP based signaling for VOIP; Networking protocols; Voice gateway for conversion of packet switching to circuit switching and vice versa; C and Perl programming; Linux IPC and socket programing; Testing methodologies: unit testing, regression testing, functional and performance (load) testing;
Embedded Software Engineer, NMS communication Corp., Schaumburg, Illinois (1998-2000)
RTOS kernel development – design, test, and improve for MC680xx based PCI/CPCI cards to support SS7 signaling and VoIP.
• Enhanced the kernel to support priority IP packets handling used by SS7 redundancy features. Designed tasks to test the new features.
• Investigated Expect script based testing platform, created testing suites to perform various functional and performance testing.
• Found hardware defects in Lucent TSI chip T8100/T8100A caused by inaccurate synchronization signals.
• Designed a task to monitor the overflow of memory stacks.
• Technologies used: RTOS kernel development; C and assembly programming; Networking protocols, such T1/E1, Ethernet, TCP/UDP/IP, SS7; Testing automation based on scripts, such as Expect, TCL/Tk;
• PhD, Electrical Engineering (signal processing), Illinois Institute of Technology, Chicago, 2004
• MS & BS, Electrical Engineering (control systems), Tsinghua University, Beijing, China
Green card holder