About This Webinar
Presenters:
Professor Stephen Boppart
The Future of Healthcare With Optical Biomedical Imaging
A medical doctor and professor at the University of Illinois, Boppart's remarks will emphasize the role of medical imaging and how this technology has enabled us to look into the body at many different size scales, how imaging has enabled disease diagnosis, and how imaging has made a difference in health care. His remarks come from a presentation he shared with members of Congress and their staffs in November on how federal support has driven the development of photonics imaging technology that is having positive impact on patients’ lives every day, a congressional briefing convened by the Optical Society of America.
Professor Boppart heads the Biophotonics Imaging Laboratory at the university's Beckman Institute and is a full-time professor in the Bioimaging Science and Technology group. As imaging trends progress to the molecular level, the origin of virtually all disease processes, the ability to image or map the location of endogenous molecules or exogenous molecularly-targeted contrast agents will become important for diagnosis. Professor Boppart's research spans a wide range of novel optical technologies, systems, methods, and applications including molecular imaging, novel contrast agents for OCT, functional OCT for imaging neural activity and physiological parameters, and multimodality optical imaging techniques for tracking the growth and development of engineering tissues, tumors, and biological specimens. He also has translated many of these optical diagnostic technologies into clinical studies with several ongoing patient trials at local medical institutions.
Questions and Answers from the Webinar for Professor Boppart:
How and when will wafer-level cameras affect the medical imaging market?
There will likely be a growing market for these cameras. A large amount of optical imaging in medicine is in using endoscopes or laparoscopes or devices to view internal body structures and surfaces. Smaller, more sensitive, and higher resolution imaging with CCD-based cameras will play important roles here.<
Could you please elaborate on the use of the optical methods you described for diagnosis and treatment of breast cancer?
In breast cancer, I described the use of OCT for intraoperative surgical guidance to assess surgical tumor margins or lymph nodes. OCT can provide microscopic imaging in real-time, compared to histological imaging on resected, sectioned, and stained sections that often takes days to process. More information and publications can be found at
http://biophotonics.illinois.edu. In other areas, optical imaging has been used to perform optical mammography using diffuse optical tomography (DOT), and more recent techniques are tracking fluorescent probes during surgery to identify lymph nodes prior to resection.
Besides current usage of semiconductor image sensors, what is your view and prediction for the new application field in medical, except OCT?
I believe the medical field is ripe for technological innovation. Radiology is a primary driver for new technology and imaging, but we are seeing electronic medical records evolve, as well as various biosensing and assay-based diagnostic technologies for personalized medicine. There are many more opportunities where technology will improve our practice of medicine and our healthcare overall.
Is there application for photonic imaging in image-guided minimally invasive surgeries and procedures?
Yes, very much so. Please see my previous responses. Current minimally-invasive surgeries and procedures rely entirely on video CCD-based imaging. I foresee that many other optical imaging technologies will be developed for use in these scenarios, such as OCT, fluorescence imaging, nonlinear optical imaging, among others.
How can the resolution of OCT be improved?
The axial and transverse resolutions in OCT are somewhat independent. The axial resolution is inversely dependent on the bandwidth of the optical source. So, broad bandwidth sources are preferred, or fast wavelength-swept sources that cover a broad spectral range. Currently, bandwidths of 100+ nm are possible, giving 2-5 micron axial resolutions. The transverse resolution is dependent on the beam focusing optics, so higher numerical apertures give higher transverse resolutions, but at the expense of a reduced depth of field. Computational approaches like our Interferometric Synthetic Aperture Microscopy (ISAM) take a computed-imaging approach to overcome the reduced depth-of-field limitation while still maintaining high transverse resolution. I believe there are still opportunities to improve OCT resolution, but we are close to the point where the technical complexity is increasing rapidly for only incremental improvements in resolution.
What is the speed of collected OCT images? For example, could I use OCT to collect a series of images looking through a mass of moving honeybees to visualize the creation of the honeycomb underneath?
OCT systems are typically acquiring data at rates of 50,000 to 300,000 axial (a-scans) scans per second. Each a-scan is one column in a 2-D b-mode OCT image. If the OCT image has 500 columns across, then frames rates are at 100-600 images per second. 3-D OCT is regularly performed, and rates are getting fast enough for real-time 3-D volumetric imaging. These rates are fast enough to capture much of the motion in moving honeybees, but imaging through a mass of bees would not be possible given the imaging depth limits of 2-3 mm in highly-scattering tissues.
How far are we from visualizing light propagating in a strongly scattering media, like brains?
OCT has been used in highly scattering tissue, such as brain tissue, to depths of about 1-2 mm (but giving micron-scale resolution). Diffuse optical tomography (DOT) is routinely used in imaging human brain activity, and these results have been correlated well with functional MRI. Because diffusing light is highly scattered and absorbed, resolution in this tomographic approach is much lower (worse) that in OCT, typically on the order of several millimeters.
Could you comment on the desire for and benefit of enhanced sensivity in the IR spectrum?
OCT and most all optical biomedical imaging techniques rely on light in the near-infrared (700~1300nm) because these wavelengths comprise the “Biological Window” in tissue where absorption is minimal and attenuation is governed most strongly by scattering. Further into the IR, light is strongly absorbed by water, and penetration in biological tissue is reduced significantly for imaging. There continues to be a strong need for enhanced sensitivity in the NIR spectrum.
Nobody mentioned the growing application and value of spectral imaging in either of these applications. Can anyone provide their opinion on this quickly emerging technology?
Hyperspectral imaging does play an important role in optical biomedical imaging, as it enables enhanced contrast in visualizing different structures such as blood vessels or contrast agents. Spectral imaging is also readily integrated into existing optical systems in endoscopes and laparoscopes, or microscopes. I believe we will see more of this advance in the future.
Kevin Harding
The Promise and Payoff of 2D and 3D Machine Vision
A 30-year optics industry veteran, Kevin Harding is an SPIE Fellow and past president of SPIE. He has served as conference chair or co-chair on 15 SPIE conferences and taught over 60 short courses industry wide on machine vision optics, 3-D imaging, laser sensors, and optomechanical technology. He has contributed over 100 papers, including 3 trade journal articles and 6 book chapters in the area of optical inspection and measurement technology, and has received 35 patents. Harding is principal engineer and project leader for the General Electric Global Research Center in New York where he leads the activity in Optical Metrology.
Questions and Answers from the Webinar for Kevin Harding:
Nobody mentioned the growing application and value of spectral imaging in either of these applications. Can anyone provide their opinion on this quickly emerging technology?
Spectral imaging is emerging, but like many higher end tools has been emerging for a while. It offers some good capability, but the capabilities of cameras and processing to data analysis has more activity around it so keeps catching up.
From the point of view of image sensors: What is their main problem to machine vision, whose solution will provoke a boost in the market? Why are the smart sensor market so little in comparison to specific ones, is it the price or other difficulties?
I am not sure I follow the question, but the challenge for machine vision regarding image sensors has always been that machine vision does not have the level of volume that say home photography or cell phones does, and so machine vision does not drive the sensor market, but rather benefits from the advances in consumer markets. Smart cameras are for practical reasons targeted to be inexpensive. Smart cameras compete with old comparators and hand gages so need to be inexpensive to work in this applications space, so tends toward less expensive systems.
Where do you see the growth in the security market? Could you please be specific?
There is a drive in the security market in such areas as tracking of people in stores to determine not only bad intent, but also product interest and trends. There is also a drive to improve biometrics to be vision based rather than contact for the benefit of both speed and the ability to have more ‘unmanned’ access points that provide high quality images, be it fingers, faces, irises or other. There is more volume in security at automated tellers than the entire machine vision industry. Now that prices are down, these volume applications will be growing.
Years ago, there were many suppliers, and the ones that survived are the ones that chose a focus area they were good at, and stuck with it. The ones that did not survive tried to take on too many areas, could not provide good expertise in a wide variety, so could not build clientele. Learning to say no, we don’t do that establishes a stronger position than trying to do anything that comes along.
Question for Kevin: How far are we from visualizing light propagating in a strongly scattering media, like brains?
This is more the field of OCT, which was developed specifically for imaging in scattering media. Other imaging methods such as synthetic aperture and holographic imaging (see for example Bahram Javidi at U of Conn) also offers potential in this space. The question is more finding what information that can provide over more conventional methods such as NMR or CT scanning.
Paul Kellett
Trends and Developments in the Machine Vision Market
Kellett will provide an overview of trends and developments in the Machine Vision market, with a special emphasis on the impact of the economy.
Paul Kellett is director of Market Analysis at the Automated Imaging Association (AIA). He has been with AIA for over eight years and performed market research and strategic marketing for over 15 years. Prior to working for AIA and its sister associations, he was president and Principal Market Analyst at Accuture Telecom Research, a firm he founded. Before his tenure at Accuture, he held the positions of senior director – Research and Senior Market Analyst at Pioneer Consulting, a Boston, MA. firm, specializing in broadband and optical networking research. Before that, he held numerous marketing, corporate planning, financial and regulatory positions at SBC/Ameritech (now AT&T). Kellett began his career at Michigan Bell as a revenue forecaster.
Questions and Answers from the Webinar for Paul Kellett:
Do you see 3D Machine Vision growing at the same rate at the general lighting values you have shown?
The drivers of demand for 3D MV and MV lighting are different, and so I would not expect these different product markets to growth at similar rates. The main driver of demand for MV lighting units in the aggregate is demand for MV cameras. (That’s not surprising, since MV cameras need lighting to function.) If you think of demand in terms of sales revenue as opposed to units, you also have to consider the downward trend in unit price, which reflects a change in the technology mix for lighting. “Technology mix” refers to the fact that LEDs have become the lion’s share of lighting sales.
The drivers of demand for 3D systems involve demand for the special applications for which 3D systems are suited (for example applications involving metrology). The specific demand drivers have to be considered within the context of individual industries.
There are so many small Machine Vision suppliers on the market. For a small supplier, how do you think we should strategize to be successful?
Essentially, you need to match your core competencies with a niche market. You need to be clear about your company’s strength and how they are different from those of your competitors. Unless you have different strengths, you won’t be able to differentiate yourself, which means you won’t be able to provide customers with a reason for doing business with you as opposed to your competitors. Those special strengths will allow you to effectively position yourself vis a vis a niche market that you identify. This means that in addition to having special strengths, you have to carve out a niche market.
If you can’t do these things, you might want to enter into strategic alliances with other companies and become part of a consortium, which has greater scale and market power. Or you might want to merge your assets with a larger company in recognition of the fact that the long-term trend is toward market consolidation.
What is the current status and future direction of international competition in the machine vision business?
Machine vision is already international. Most firms, large and small, export all over the world. Germany, Japan, the U.S. and Canada have most of the MV companies. The larger suppliers are more apt to have in-country subsidiaries, while smaller companies rely primarily on domestic distributors. Of course the ability to export to specific companies is greatly affected by exchange rates and trade tariffs and procedures. As time goes on, I would expect markets to become more open to imports and thus competition will become even more intense.
From the point of view of image sensors: What is their main problem to machine vision, whose solution will provoke a boost in the market? Why is the smart sensor market so little in comparison to specific ones, is it the price or other difficulties?
The area chart I showed did indicate that ASMV systems are a large product market in terms of dollars (sales revenue) than smart cameras. But keep in mind that the chart would be far less lop-sided in favor of ASMV systems than smart cameras, if I had based it on units. The reason is that ASMV systems cost more ($20 to $60 thousand USD), while smart cameras cost $2 to $10 thousand. In actuality, smart camera sales are growing faster than ASMV systems.
Nobody mentioned the growing application and value of spectral imaging in either of these applications. Can anyone provide their opinion on this quickly emerging technology?
Spectral imaging is a very specific technology in its own right. Our presentations were very high-level and designed to give the “lay of the land” from 20 thousand feet up. Moreover, demand for products based on spectral imaging has not yet been researched to the best of my knowledge. But it stands to reason, that if spectral imaging can (a) satisfy certain customer needs better than other technologies, (b) do so cost-effectively and (c) address a sufficient number of customers to generate sufficient revenue, then spectral imaging could indeed be very promising.