Date of Award

1998

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Computer Science

First Advisor

John M. Tyler

Abstract

Set compression allows the compression a set of similar (correlated) images more efficiently than compressing the same images independently. Currently, set compression is performed with different inter-image predictive models, that forecast the common image properties from a few reference images. With sufficient inter-image correlation, one can predict any database image from a few templates, hence avoiding inter-image redundancy and achieving much improved compression ratios. This research focused on two major aspects of this technique: the practical limits of the predictive set compression, and the theoretical estimates of the compression efficiency. This includes a review of the previous work in set compression area, a discussion of the more important statistical and informational aspects involved in predictive set compression, practical observations and measurements for medical (CT and MR) data, and theoretical analysis of lossless similar image compression. This research proposes new and more reliable approaches for lossless set compression, as well as their extensions to more general lossy set compression.

ISBN

9780591997989

Pages

147

DOI

10.31390/gradschool_disstheses.6755

Download

Share

COinS