OBJECTIVE

Accomplished IT professional with a proven track record seeking a full-time position in field of image and video processing, computer vision, computer graphics, machine learning and pattern recognition.

SUMMARY [Back to top]

• Innovative problem solver specialized in algorithm design and optimization, with 7 patents related to machine learning, pattern recognition and image processing.

• Result-driven developer with 4 years' experience in the design, development and deployment of commercial software, with expertise in C/C++, C# and MATLAB.

• Extensive experience developing cross-platform software in C++, with deep understanding of object-oriented design.

• Published author and conference presenter in field of image processing and pattern recognition.

• Ph.D. degree in Electrical and Computer Engineering.

WORK EXPERIENCE [Back to top]

Senior Software Engineer and Imaging Architect                              November 2006 ~ Present The Neat Company, Inc., Philadelphia, PA

• Working as the imaging architect, built a cross-platform (Windows, Mac and Linux) imaging library. Designed and implemented all advanced image processing algorithms. The library successfully removed the dependency on a commercial software, reduced the product cost and saved tremendous customer support time. Highlighted components include:

  An adaptive image binarization algorithm. The algorithm specializes in handling low-quality images by dynamic parameter decision. Most importantly, it improves the performance in both processing time and image quality.

  An adaptive image enhancement algorithm. The algorithm successfully solved the image quality problem that many customers used to complain about. The enhancement of one large color image (2500?3300) usually takes only 200 milliseconds; and the file size can be reduced by 20%~50%.

  An image segmentation algorithm. Applied to any flatbed scanner, the algorithm can save at least 50% of scanning time by automatically detecting and separating objects in an image.

  (Patents have been filed for most of the work.)
  

• Working as the algorithm and development lead, designed and delivered a variety of intelligent systems by successfully applying image processing and pattern recognition technologies in applications. Created new features for the company's products. Major contributions include:

  An image-based document classification engine. The engine extracts physical image features (page orientation, margin smoothness, colorfulness, logos and so on) to carry out automatic classification and can achieve accuracy as high as 94%.

  An invoice processing engine. The engine separates an invoice image into graphics, barcode, table grid, stamp and text layers. The layer information is combined with OCR results to intelligently parse important invoice information.

  A logo detection and recognition engine. The engine significantly boosted the performance of both document classification and information parsing.

  (Patents have been filed for some of the work.)

• Working as the senior software engineer, developed and delivered a cross-platform (Windows and Mac) and thread-safe Software Development Kit (SDK) that provides C++, C# and Objective-C APIs. The SDK enables other developers to create high value and competitive products while minimizing implementation time. It also brought about the following new product features:

  Blank page detection engine. It enables automatic detection and removal of blank pages to save space and editing time.

A multiple image acquiring engine. The engine applies multi-thread technique to enable simultaneous image capture on multiple image acquiring devices.

• Designed and implemented a NUnit-based testing suite. The testing suite embraces all functionalities in the SDK to perform both smoke and stress testing. Working with the SVN server, testing is automatically triggered by new code changes to ascertain code functionality.

(Used C++, C#, STL and Boost and applied design pattern.)

Research Scientist                                                                              February 2006 ~ October 2006
Applied Media Analysis, Inc., College Park, MD

• Designed a mobile business card reader. Optimized all image processing and recognition algorithms for mobile devices. The card reader turns a camera phone into a powerful contact management tool and usually takes less than 10 seconds to process a card image.

• Designed and optimized all algorithms required by a mobile barcode reading system. The system is able to use a mobile device to capture, detect, distinguish and decode eight different types of barcodes.

• Designed a mobile road sign capture and segmentation system. The system uses a mobile device to capture a road sign image and segment texts on it.

(Used C++ and worked on both Windows Mobile and Symbian systems.)

Software Engineer, Summer Intern                                                        June 2005 ~ August 2005
Applied Media Analysis, Inc., College Park, MD

• Designed an Arabic keyword spotting algorithm and optimized it for mobile devices. Arabic words on device-captured images are encoded with shape descriptors. Word spotting is carried out using approximate string matching. The designed algorithm is able to accurately locate a target word in less than 3 seconds on a mobile device. (Used C++.)
  

EDUCATION [Back to top]

Electrical and Computer Engineering, Ph.D.                                                                     2005
University of Maryland, College Park, MD

Dissertation: "Adaptive Analysis and Processing of Structured Multilingual Documents". The research focused on new machine learning and pattern recognition algorithms that accomplish rapid information extraction from multilingual documents. Designed a new bootstrapping algorithm that enables intelligent structure analysis of bilingual documents. Also designed a re-targetable learning algorithm that helps rapid creation of new language OCR systems. The research has been applied to cross-language information retrieval and machine translation systems, and has provided great convenience to English speakers to search and interpret critical information in new languages without requiring any knowledge of those languages. (Used C++ and MATLAB.)

Electrical Engineering, Ph.D. Study                                                                                     2000
University of Hawaii, Honolulu, HI

Worked on the research topic "3D Image Registration". Designed a new methodology to generate range images from volume data based on ray projections. Triangular meshes at different error levels were then created for each generated range image. A progressive algorithm was implemented to conduct the 3D registration, where coarse meshes were used to obtain the initial result and fine meshes were used to refine the result. (Used C++, OpenGL and VRML.)

Electrical Engineering, M.S.                                                                                                    1996
Beijing Institute of Technology, Beijing, China

Electrical Engineering, B.S.                                                                                                    1993
Northwestern Polytechnical University, Xi'an, China

PATENTS [Back to top]

1. H. Ma, "Table Grid Detection and Separation" (Filed in 08/2009)

2. H. Ma, "Image Background Enhancement for Color Images" (Filed in 08/2009)

3. H. Ma, "Method and Apparatus for Binarization Threshold Calculation" (Filed in 08/2009)

4. H. Ma, "Method and Apparatus for Despeckling an Image" (Filed in 08/2009)

5. H. Ma, "Method and Apparatus for Calculating the Background Color of an Image" (Filed in 08/2009)

6. H. Ma, ?Method and Apparatus for Modifying the Histogram of an Image" (Pending)

7. A. Turkelson, H. Ma, "Automatic Document Classification Using Lexical and Physical Features" (Pending)

SKILLS [Back to top]

Languages: C/C++, C#, Objective-C, Cocoa, MFC, OpenGL, DirectX, Java, Pascal, Fortran, Basic, Assembly, STL

PUBLICATIONS [Back to top]

1. M. Agrawal, H. Ma and D. Doermann, "Chapter: Generalization of Hindi OCR using Adaptive Segmentation and Font Files", Guide to OCR for Indic Scripts, Springer, 2009.

2. S. Jaeger, H. Ma and D. Doermann. "Combining Classifiers with Informational Confidence". Studies in Computational Intelligence: Machine Learning in Document Analysis and Recognition, pages 163-192. Springer, 2008.

1. Z. Wang and H. Ma, "Advanced Continuous Wavelet Transform Algorithm for Digital Interferogram Analysis and Processing", Optical Engineering, Vol. 45, No. 4, 2006.

2. H. Ma and D. Doermann, "Adaptive Hindi OCR Using Generalized Hausdorff Image Comparison", ACM Transactions on Asian Language Information Processing, Vol. 26, No. 2, 2003, pp198-213.

3. H. Ma, B. Karagol-Ayan, D. Doermann, D. Oard and J.Wang, "Parsing and Tagging of Bilingual Dictionaries", TAL Traitement Automatique Des Langues, Vol. 44, No. 2, 2003, pp125-150.

1. H. Ma and D. Doermann, "Adaptive OCR with Limited User Feedback", 8th Int'l Conf. Document Analysis and Recognition (ICDAR), 2005, pp814-818.

2. S. Jaeger, H. Ma and D. Doermann, "Identifying Script on Word-Level with Informational Confidence", 8th Int'l Conf. Document Analysis and Recognition (ICDAR), 2005, pp416-420.

3. Z. Wang and H. Ma, "Automatic analysis of photomechanics interferogram using wavelet transform", Proc. SEM Annual Conference and Exposition on Experimental and Applied Mechanics, 2005.

4. H. Ma and D. Doermann, "Font Identification of Scanned Documents Based on Texture Features Using a New Texture Operator", SPIE Conf. Document Recognition and Retrieval, 2005, pp148-156.

5. H. Ma and D. Doermann, "Adaptive Word Style Classification using a Gaussian Mixture Model", Int?l Conf. Pattern Recognition (ICPR), 2004, pp606-609.

6. H. Ma and D. Doermann, "Application of Three Classifiers to Word Level Script Identification on Scanned Document Images", SPIE Conf. Document Recognition and Retrieval, 2004, pp124-135.

7. H. Ma and D. Doermann, "Gabor Filter Based Multi-class Classifier for Scanned Document Images", 7th Int'l Conf. Document Analysis and Recognition (ICDAR), 2003, pp968-972.

8. H. Ma, B. Karagol-Ayan, and D. Doermann, "Segmenting and Tagging Structured Content", Symposium on Document Image Understanding Technology (SDIUT03), 2003, pp53-64.

9. H. Ma and D. Doermann, "Bootstrapping Structured Page Segmentation", SPIE Conf. Document Recognition and Retrieval, 2003, pp179-188.

10. D. Doermann, H. Ma, B. Karagol-Ayan and D. Oard, "Lexicon Acquisition from Bilingual Dictionaries", SPIE Conf. Document Recognition and Retrieval, 2002, pp37-48.

11. H. Ma, T. Lu and D. Y. Y. Yun, "Progressive Registration of Range Image Based on Surface Triangulation", SPIE's Int'l Symposium on Optical Science and Technology, 2000, pp79-80.
    

Technical Reports

1. H. Ma and D. Doermann, "Adaptive Hindi OCR Using Generalized Hausdorff Image Comparison", LAMP-TR-105, CAR-TR-987, CS-TR-4519, UMIACS-TR-2003-87, University of Maryland, August 19, 2003.

2. H. Ma, B. Karagol-Ayan, D. Doermann, D. Oard and J. Wang, "Tagging and Parsing of Bilingual Dictionaries", LAMP-TR-106, CAR-TR-991, CS-TR-4529, UMIACS-TR-2003-97, University of Maryland, September 19, 2003.

PROFESSIONAL SERVICES [Back to top]

International Conference on Document Analysis and Recognition, International Conference on Pattern Recognition and International Conference on Image Analysis and Processing.

Contact Me [Back to top]

EMAIL: my first name AT gmail DOT com

Telephone: Please send email to request.