Certain image processing functions can be implemented more efficiently when the input data is in compressed form. Such an experimental system has been studied and simulated. The system consists of a one-dimensional Differential Pulse Code Modulation (DPCM) compressor, a one-dimensional non-recursive linear filter, and a one-dimensional DPCM decompressor, applied in that order. The implementation is more efficient because the filter is applied to the data in their compressed form, where fewer bits per pixel are required to represent them. A second, more conventional, system that contains the same functional elements but reverses the order of the filtration and decompression operations has also been implemented for comparison to the experimental one. The differences (errors) between the signals output from the two systems have been modeled and the models validated through experiments. It has been found that the systems can be made to yield equivalent results if certain parameters are constrained. These constraints do not put undue demands on system design nor do they substantially degrade system performance. Images produced by the two systems are presented and suggestions for additional work are discussed.
Library of Congress Subject Headings
Image processing--Digital techniques
Department, Program, or Center
School of Photographic Arts and Sciences (CIAS)
Masia, Andrew, "Digital processing of compressed image data" (1984). Thesis. Rochester Institute of Technology. Accessed from
RIT – Main Campus