Proc. of 5th World Multiconference on Systemics, Cybernetics and Informatics (ISAS-SCI 2001), Orlando, FL, July 22-25, 2001.
Stephen Wolf; Margaret H. Pinson
Abstract: This paper presents objective-to-subjective correlation results for a reduced-reference, in-service, video quality monitoring system. This reduced-reference system utilizes quality parameters that are computed by comparing features extracted from spatial-temporal (S-T) regions of the input video stream with identical features extracted from the output video stream. The amount of reduced-reference information that is required to compute the quality parameters is inversely related to the size of the S-T region. Smaller amounts of reference information (i.e., larger S-T regions) are desired since less transmission or storage bandwidth is required for the reference information. However, objective-to-subjective correlation drops off if the S-T region size becomes too large. In this paper we examine the tradeoffs between objective-to-subjective correlation results and S-T region size. Correlation results for S-T region sizes from 8 vertical lines x 8 horizontal pixels x 2 video frames to 128 x 128 x 24 are presented. These results utilized a total of nine subjectively rated data sets that span an extremely wide range of bit rates and compression techniques. Thus, designers of television video systems as well as Internet video streaming systems may use the results.
Keywords: objective; subjective; correlation; reduced-reference; in-service; video quality metrics; spatial; temporal; compression; sub-region
For technical information concerning this report, contact:
Margaret H. Pinson
Institute for Telecommunication Sciences
Disclaimer: Certain commercial equipment, components, and software may be identified in this report to specify adequately the technical aspects of the reported results. In no case does such identification imply recommendation or endorsement by the National Telecommunications and Information Administration, nor does it imply that the equipment or software identified is necessarily the best available for the particular application or uses.