Research Repository

Blockiness Estimation in Reduced Reference Mode for Image Quality Assessment

Qadri, MT and Ghanbari, M (2014) 'Blockiness Estimation in Reduced Reference Mode for Image Quality Assessment.' Bahria University Journal of Information & Communication Technologies, 7 (1).

BUJICT14_Issue_Paper-7.pdf - Published Version

Download (1MB) | Preview


In this paper we present blockiness estimation meter in reduced reference mode using frequency domain analysis. Since blockiness is an abrupt luminance discontinuity at the DCT block boundaries and also has periodic patterns, therefore blockiness generates harmonics in frequency domain. In order to design the quality meter in reduced reference mode, we used the strength of harmonics (of reference image) as reduced reference parameter and compared it with the harmonics of the coded image to estimate blockiness distortion. The severity of blockiness artifact depends upon the strength of harmonics generated in the coded image. The spatial masking is also applied in order to mask the distortion according to the local spatial activity of the image. As the distortion is content specific therefore it is calculated locally for each part of the image by applying block processing technique. To avoid misinterpreting the natural edges lying at the DCT block boundaries, we also applied the edge cancellation process which distinguishes the natural edges and the edges due to blockiness in the coded image. Finally the designed algorithm is tested on the LIVE image database which consists of 233 colored images compressed at different compression rates and when tested with the database, the correlation coefficient of 93.59% is achieved in reduced reference mode.

Item Type: Article
Uncontrolled Keywords: Blockiness estimation, reduced reference, frequency domain analysis, image quality estimation
Divisions: Faculty of Science and Health
Faculty of Science and Health > Computer Science and Electronic Engineering, School of
SWORD Depositor: Elements
Depositing User: Elements
Date Deposited: 31 Jul 2019 08:46
Last Modified: 06 Jan 2022 13:51

Actions (login required)

View Item View Item