Author
Ozansoy, Koray, Özer, N., Dönmez, F., Uğurdağ, Hasan Fatih
Publication Date
2014
Publication Place
-
IEEE
Subject
Data compression, Edge detection, Image processing
Type
Document
Language
English
Digital
Yes
Manuscript
No
Library
Özyeğin University
Library Asset ID
2165-0608
Record ID
16c987fc-42a5-4c93-b508-f9090ba8c3c0
Library Location
Electrical & Electronics Engineering
Date
2014
Notes
Due to copyright restrictions, the access to the full text of this article is only available via subscription.
Sample Text
In today's world where video traffic is at its peak on the Internet and in service provider networks, the benefits of automatic image quality measurement are obvious. These measurements need to be made in real time in many applications, and this requires measurement without the use of "No Reference", that is, uncompressed (raw) images. The majority of video streams in the world are now digital. Digital video streams use compressed video transmission, and most of the compression methods used are DCT-based. In such streams, the decrease in image quality usually occurs due to the transmission speed being reduced too much, and since the DCT algorithm is block-based, the loss of quality manifests itself as "Blocking". In this study, we present a blocking measurement method (called RED) that gives results closer to human perception than the methods in the literature. One of the most important contributions of RED is that it has managed to establish an analytical relationship between the blocking values it automatically calculates and the grades given by testers. RED optimizes the parameters of this relationship with “Regression”. Again, unlike the literature, RED excludes some blocks from the calculation by using "Edge Detection" before calculating the blocking.
DOI
10.1109/SIU.2014.6830679
