How do you think the new GigE standards will influence the machine vision industry?
Respond or ask your question now!
By Lee J. Nelson
Dr. Levine: The main emphasis of most academic research was on so-called low-level processing prior to 1986. Most “inexpensive” computers were defined by what academicians thought they could use or needed to test their theories. This was particularly true for the potential application of sophisticated methodologies to interesting, real-time imaging problems.
Mr. Schatz: Machine vision technology was in a “sorry” state in 1986. Literally hundreds of vision companies were chasing any and every conceivable application. However, the products delivered at that time rarely met customer expectations for robustness and reliability. This was due to a combination of technological and business factors: we depended either on an underpowered general purpose computer platform or overpowered specialty hardware. Also, most systems had a clumsy user-interface. Vision companies underestimated the challenges of making products for an industrial environment, while customers had the expectation that expensive systems actually would work as well in their factory as they did at demonstration.
AI: Describe the changes you’ve seen in imaging over the past two decades?
Mr. Biegel: Display methodology has evolved from film and videotape to distributed softcopy. Exchange of medical image data via DICOM [Digital Imaging and Communications in Medicine] now is a given. Transport of metadata and workflow integration still are a challenge, but there is good progress on a number of fronts. Development of underlying standards, such as DICOM and work in the IHE [Integrating the Healthcare Enterprise] committees, are solid efforts and key to reducing medical errors while improving clinical workflow solutions. Another major change/trend is the explosion of imaging-related information used in routine clinical practice. The evolution of image-enabled care is causing diagnostic imaging not only to transcend simple data capture, but to extend into the daily routines of referring physicians and other health care team members.
Mr. Gibson: During the last two decades, there have been major changes in sensors, optics, camera integration and software control. CCDs, for example, first moved to interlined sensors, then to single-channel and more recently to dual-channel CMOS devices. Mass produced glass optics have made it easier to improve light columniation. As further improvements were made in composite materials, quality optics could be had at lower prices. With improvements in chip power management—as well as the ability to embed significant algorithm support inside ICs—the expertise to implement imaging research and development programs is focusing more on real-world applications and less on the pure science.