1 February 1991 Legibility of compressed document images at various spatial resolutions
Author Affiliations +
Abstract
Digital images of business and financial documents are replacing paper and microfilm in numerous applications. System bandwidth limitations usually necessitate data-compression of such images. Another way to decrease the bandwidth requirement for each image is to reduce the spatial resolution of the image, thereby reducing the total number of pixels which must be stored or transmitted. In this paper, the authors investigate the combination of spatial resolution reduction with the evolving ISO/JPEG grayscale image-compression standard. Both resolution reduction and ISO/JPEG data-compression cause distortion of the decompressed image with respect to the original. The system designer has some flexibility in allocating this distortion to each of the processes. The authors have conducted extensive testing to determine the best such allocation. Results of these tests indicate that legibility depends principally on compressed size. In particular, legibility is directly proportional to compressed size. The effects of spatial resolution on legibility are secondary. One such effect is a decrease in legibility with increasing spatial resolution as packet size is held constant at comparatively small values. Taken together, these results argue that the best legibility for document images compressed with the ISO/JPEG draft standard is obtained at the lowest adequate spatial resolution.
© (1991) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Robert C. Kidd, Robert C. Kidd, Charles Martin Zalinski, Charles Martin Zalinski, Jerome I. Nadel, Jerome I. Nadel, Robert D. Klein, Robert D. Klein, } "Legibility of compressed document images at various spatial resolutions", Proc. SPIE 1454, Beam Deflection and Scanning Technologies, (1 February 1991); doi: 10.1117/12.28025; https://doi.org/10.1117/12.28025
PROCEEDINGS
11 PAGES


SHARE
Back to Top