John F. McGowan, Ph.D. is a Scientist with GFT Group Incorporated, a private research and development company. He was a project manager and key contributor for CompCore Multimedia's SoftDVD product, one of the first Microsoft Windows DVD players, as well as other CompCore products. He has developed software implementations of image quality metrics, perceptual optimization of JPEG image compression, and video quality metrics for NASA. He has contributed to the conceptual design of systems for televising missions to Mars for NASA.
Dr. McGowan offers the following services primarily in the area of digital video, audio, imaging, and multimedia for entertainment, communications, security, biomedical, and other applications:
These services are described in detail below.
An algorithm is a predetermined set of instructions for solving a specific problem in a limited number of steps. Computers and biological systems rely on sophisticated algorithms to perform speech recognition, vision, image and video compression, blood clotting, and many other tasks.
Rapid advances in semiconductor processing technology (Moore's Law) have resulted in the equivalent of a traditional supercomputer on every desktop in industrial nations. These advances have also produced inexpensive digital cameras, video cameras, audio and video capture cards for personal computers, and desktop image scanners. Most of these computers are under-utilized due to the lack of algorithms that can exploit their processing power for useful applications such as computer vision and speech recognition.
Saturation of the office automation markets means that the computer and networking industries need new products for continued sales growth. New computational algorithms or new uses of known computer algorithms offer a solution to this problem.
Compact, light-weight, low power video surveillance and reconnaissance systems for business and home security, national security, news gathering in hazardous environments, and other applications are the most immediate example of this next wave of technology. Such systems already exist and can be implemented using MPEG or H.26x audio and video compression technology and compact multi-gigabyte hard drives or other storage technology. The entire system will fit in the palm of a man's hand and weigh less than a kilogram (2.2 pounds). The terrorist attack of Sept. 11, 2001 should give a big boost to the market for these systems. Studies of compact, light-weight, low power video systems for Mars exploration, a much more demanding application than most terrestrial applications including terrestrial military applications, can be found here and here.
Dr. McGowan offers a number of services to enable companies to capture these new markets.
General solutions for problems such as computer vision and speech recognition appear quite difficult. To exploit these markets and technologies, companies need to identify immediate markets for partial solutions - and in some cases even for complete solutions - to these problems.
Algorithm research and development is unpredictable. The capabilities of the algorithms that may be developed are difficult to anticipate in advance and cannot always be dictated by customer requirements. Thus companies need to identify a spectrum of potential markets and products both to guide research and development and to maximize the probability of exploiting any technical advances.
Dr. McGowan offers services to identify potential markets, define products that can exploit these markets, and specify the technical capabilities that an algorithm will need to exploit these markets.
New technologies such as multimedia, speech recognition, and computer vision rely on sophisticated algorithms such as the Discrete Cosine Transform (MPEG), wavelets (JPEG-2000), artificial neural networks (computer vision), and hidden Markov models (speech recognition). Major advances in these areas are likely to use currently unknown or unrecognized mathematical methods. Investors and companies have invested hundreds of millions, even billions, of dollars in companies and projects that were not technically feasible, such as Lernout and Hauspie in speech recognition and GO in handwriting recognition, and that often simply recycled known technologies with serious practical limitations.
Scientists and technologists frequently present problems such as speech recognition and handwriting recognition as solved or nearly solved, except for a few technical details. Technology demonstrations and prototypes often utilize known algorithms that can be impressive in a controlled demonstration but do not work in the real world and that may be scientific dead ends. This has resulted, in some cases, in large wasted investments and the diversion of research and development funds away from novel, unorthodox approaches that actually have a much higher likelihood of success.
Dr. McGowan offers independent technology assessment of computational algorithms for due diligence and other purposes.
Estimating the cost and schedule and effectively managing product development for products depending on complex algorithms can be difficult due to the complex and advanced nature of the underlying technology. Most often, resource requirements and costs are underestimated. For example, the difficulty of porting complex algorithms implemented in C or C++ from one processor, e.g. the Intel Pentium, to another, e.g. the PowerPC, is frequently underestimated. In some cases, the technical difficulty of the project can be overestimated because the technology involved seems impressive to the uninitiated but is actually easy to implement or to acquire from a third party.
Dr. McGowan offers project management services for software development projects and research and development projects using advanced computational algorithms such as MPEG digital video.
Technical standards such as the ISO MPEG digital video standard and the ITU H.26x video standards are quite complex. In some cases, the organizations and companies that developed and established the standards have deliberately or accidentally exploited subtle technical aspects of the standards to create de facto technical barriers to competitors. These barriers can be both formidable and difficult to anticipate or identify.
Dr. McGowan can help identify these technical barriers and develop strategies to circumvent them if possible.
Computer vision, speech recognition, and other sophisticated algorithms that have potential multi-billion dollar markets are unsolved problems in computer science. They are unlikely to be solved by fine tuning or further incremental development of known algorithms. New mathematical approaches and "out of the box" thinking is probably required.
Dr. McGowan offers new mathematical approaches and creative "out of the box" thinking.
Future advances in medicine may require smart drugs or systems of drugs that cure or treat disease by performing simple calculations. For example, cancer cells often have abnormal numbers of biological structures such as chromosomes, centrosomes, and surface antigens. A smart drug or system of drugs that can count biological structures and kill cells with too many or too few of a biological structure such as a chromosome may provide a broad spectrum cure or treatment for many kinds of cancer.
Dr. McGowan offers the expertise to identify and develop algorithms that can attack specific diseases. He has investigated the application of various computational algorithms to problems in biology and medicine.
Dr. McGowan can be reached at:
Home Page: http://www.jmcgowan.com/