Some Detailed Descriptions of Experience:
- 1983 - present: At Northeastern U., teaching and research in imaging,
AI, pattern recognition, 3D graphics, internet webpage design. Have advised numerous
grad students including 4 Ph.D. thesis, using various languages/systems,
including UNIX, PC, Mac, JAVA, C, C++, Sybase, IE/Netscape, Windows, Ultra Sparcs,
Publication record includes over 120 technical papers, 18 books, 3 USA/Germany
patents, and awarded many grants totaling more than one million USDs.
- 2000 - 2003: the PI (Principal Investigator) of a 1/2 million grant, supporting 3
Ph.D. students, doing 3D imaging analysis, using internet/JAVA/C, PC-based systems, for
analyzing, understanding, interpreting various MRI image data, segmentation, resolution
enhancement, skeletonization, feature extraction, and classification. This can help
doctors do diagnosis for various serious illness, such as brain cancer, lung cancer,
stomach abnormal cells detection, colon cancer diagnosis etc. Our research can help
doctors do diagnosis on-line through web, with fast speed, and more than 87%
accuracy rate so far (still improving). The goal is for 99% accuracy rate.
- 9/1990 - present: at MIT Sloan School, in charge of OCR (optical character
recognition) project, in applications to bank check analysis and recognition.
I created an OCR Lab, lead a team of about half of dozen excellent researchers and
grad students, established a PC based system with scanner, using C and C++ languages,
that can scan bank checks and determine "courtesy amount" on the upper right corner",
which is the most important portion of a check that really counts. It includes machine
printed fonts, as well as handprinted/handwritten alpha-numerals. For more than 5,000
checks tested in the database, the preliminary recognition rate is over 98%
(using neural net and thinning[skeletonization] and rethichening self improving
learning algorithms). Our research results got USA patent and were published in
well-known Int. Journals (pls refer to Publication details).
- 9/1989 - 9/1990: at MIT AI Lab, doing advanced research in 3D object recognition and
stereo computer vision, multi-media vision/animation, using LISP1.5 and high
speed parallel computer with 64,000 CPU running simultaneously. Applied to industrial
parts inspection, such as pairs of scissors, hammers, boxes, etc.
- 9/1986 - 9/1996: at Harvard Univ. Extension School, Adjunct Faculty, mainly teaching
grad computer science course for industry/business professionals, including
A.I., Pattern recognition, Logic Programming, PROLOG language, Applied Automata
Theory, and Knowledge Engineering.
- 1/1980 - 9/1981: at GTE Lab, Senior Technical Staff, doing OCR (optical character
recognition) with applications to telephone communication. Leading a team of about 8
professionals (all computer engineering MS and Ph.D.'s), built a system using C
language, to analyze and recognize English characters, both handwritten and handprinted.
Designed methodologies combining advantages of structural and statistical methods,
16-vector Freeman chain coding, and a Dynamic word database, plus touch sensitive
tablet/pen, connected to computer and telephone, so that people can communicate with
both voice and written words. Prototype's accurate rate was about 89%, with large
database of about 65,000 real samples.
- 9/1981 - 9/1983: Wang Labs, Senior Software Specialist, Lead a group of 4 or 5 senior
engineers, establish an imaging system that can scan images, enhance high
resolution, and automatically integrated into office system, as part of a large
"Office Automation" System. Also done OCR work similar to the experience as in the
GTE Lab. One of my three patents was granted by the USA and Germany bureau during
this period.
Continue
Previous
Main