Inspection and Security Imaging
The Future Projects that LIVE is targeting are problems of great significance for
which there currently exist no adequate solution. One of the great problems in image
science that we are confronting is developing methods for visual screening of
potential threats in large-scale crowd settings, such as airports, outdoor and sporting
events, political rallies, and so on. We regularly hear in the news how, while our
technologies are advancing rapidly, so are those of sinister actors interested in
destruction and terrorism using weapons of various degress of lethality.
A major problem with penetrating imaging technologies is their lack of perceptual picture quality. This is true of infra-red (IR or thermal), x-ray, and millimeter-wave images, all of which often present with different degrees and types of distortions including various blur, noise, saturation, and compression processes. We are working with NIST towards modeling IR, x-ray and MM-wave picture quality, with the current aim being to be able to assess the efficiacy of a given imaging device, which is among the goals of NIST generally. However, as with the dermsocpy problem, we wish to go beyind, and develop models and algorithms to actively assist the screening expert to be able to locate weapons of all types, by introducing “quality-aware” statistical modeling into the process. It is hard to gauge the ultimate impact of this world given what little we know of developments by terrorists, but in our view, success in this direction could be enormously important. Some of our early efforts on the modeling side are found here:
T.R. Goodall, A.C. Bovik and N.G. Paulter, Jr., “Tasking on natural statistics of infrared images,” IEEE Transactions on Image Processing, vol. 25, no. 1, pp. 65-79, January 2016.
P. Gupta, J.L. Glover, N.G. Paulter Jr. and A.C. Bovik, “Studying the statistics of natural x- ray pictures,” ASTM Journal of Testing and Evaluation, to appear.