Commercialization of spectral imaging for color reproduction will require the identification and traversal of roadblocks to its success. Among the drawbacks associated with spectral reproduction is a tremendous increase in data capture bandwidth and processing throughput. Methods are proposed for attenuating these increases with data-efficient methods based on adaptive multi-channel visible-spectrum capture and with low-dimensional approaches to spectral color management. First, concepts of adaptive spectral capture are explored. Current spectral imaging approaches require tens of camera channels although previous research has shown that five to nine channels can be sufficient for scenes limited to pre-characterized spectra. New camera systems are proposed and evaluated that incorporate adaptive features reducing capture demands to a similar few channels with the advantage that a priori information about expected scenes is not needed at the time of system design. Second, proposals are made to address problems arising from the significant increase in dimensionality within the image processing stage of a spectral image workflow. An Interim Connection Space (ICS) is proposed as a reduced dimensionality bottleneck in the processing workflow allowing support of spectral color management. In combination these investigations into data-efficient approaches improve two critical points in the spectral reproduction workflow: capture and processing. The progress reported here should help the color reproduction community appreciate that the route to data-efficient multi-channel visible spectrum imaging is passable and can be considered for many imaging modalities.
Library of Congress Subject Headings
Imaging systems; Image processing--Digital techniques; Color computer graphics; Color vision; Spectrum analysis
Department, Program, or Center
Chester F. Carlson Center for Imaging Science (COS)
Rosen, Mitchell, "Navigating the roadblocks to spectral color reproduction: data-efficient multi-channel imaging and spectral color management" (2003). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus