Advanced Imaging


Advanced Imaging Magazine

Updated: July 8th, 2008 05:26 PM CDT

Challenges Change The Landscape



March 2004


Image segmentation, motion estimation and multi-resolution analysis are crucial technologies for a myriad of applications. Although they top many lists for resolving computer vision object detection and tracking applications, a comprehensive analogy can be established correlating to the massive investments being made in security applications? a.k.a. Homeland Security. The applications are widespread, for the tools needed to implement the technologies of active contours, color segmentation, camera motion estimation, density matching, Hausdorff transforms, neural networks, optical flow and even support for vector machines extend into such emerging markets as intelligent video surveillance and video mining.

While a host of US-based companies sees potential opportunities overseas for these technologies in countries such as China , there is a tsunami of activity that belies conventional wisdom. With widescale electronic R&D activities for commercial electronic imaging applications ?to include both machine vision and computer vision, in South Korea and elsewhere?as reported by the ROI Marketing Group, plus the news that China will begin fabricating semiconductors via Taiwanese expertise, this questions the conventional wisdom of how big the opportunities are for North American companies.

Since we are now wired into an electronic media, Open Source code is widely available to allow technical issues to be analyzed and resolved among peer groups. For example, color segmentation can be analyzed from shading models, active contours can be used interactively for setting up control points and deformable templates can be used for parametric models. What's more, this research activity is worldwide from India to South Korea to China .

Access to such development tools as JavaAI, Java3D, OpenGL, Intel OpenCV, Microsoft DirectX, Matlab, SVMlight and ANNIE (Artificial Neural Network for C++) are changing the development environment. Should North American and European companies be concerned about these developments? Apparently, there are many debates on this topic within both government and industry. For example, at IDGA's Image Fusion Conference (for which Advanced Imaging was the media sponsor) held in Arlington, VA, January 27 and 28, 2004, Dr. Darrel G. Hopper, principal engineer at the Air Force Research Lab, noted that display technology in the last five years has lagged behind in advancements compared with CPU, memory, hard disk and communication technologies?as much as three orders of magnitude. DoD science and technology goals for displays have ambitious plans for such technology as 300 Mpixel ultraresolution devices, true 3D holographic, volumetric multiplexed battlespace displays and intelligent displays. One has to wonder where they would get the support from the display industry since most commercial product activity is overseas. For military core display technology requirements, there is reflective LCD and H-PDLC, Flat Panel Autostereoscopic 3D, High-Definition DMD projection, High-Efficiency AMLCXD and plastic FPDs, field emissive device technology, light piping and quantum cavity displays, inorganic and organic color EL and AMEL and solid-state laser projectors.

However, Dr. Hopper makes strong comparisons for evaluating DoD's domestic display strategy. Since display improvements are absolutely necessary for all six of DoD goals, he believes the U.S. is still better at innovation in display technology. Although the U.S. defense industry relies on Asian FPD manufacturers for long-term viability, he adds, these same Asian display device manufacturers rely on U.S. technology for revolutionary innovation. Moreover, DoD's science and technology investment is about 3% of global display research, but virtually all foreign research is evolutionary from one 18-month product to the next.

Len Yencharis


Subscribe to our RSS Feeds