Use this URL to cite or link to this record in EThOS:
Title: Blockiness and blurriness measurement for video quality assessment
Author: Qadri, Muhammad Tahir
ISNI:       0000 0004 2741 7690
Awarding Body: University of Essex
Current Institution: University of Essex
Date of Award: 2012
Availability of Full Text:
Access from EThOS:
Rapid growth of wired and wireless multi-media data challenges the researchers to develop an efficient objective assessment meter to estimate their quality. This research is focused to design objective quality assessment meters for images and video sequences. In this work, we proposed blockiness and blurriness distortion meters using full, reduced and no reference approaches. Both blockiness and blurriness distortions are calculated using frequency domain analysis. Since blockiness is an abrupt luminance change at DCT block boundaries which is also periodic therefore it generates harmonics in frequency domain, we used the strength of harmonics to estimate blockiness distortion. The more severe the blockiness is, the stronger will be the strength of harmonics in frequency domain. While blurriness removes the sharpness of the image by removing the high frequency components therefore we studied the loss of high frequency coefficients to measure blurriness artefact. We also aimed to design a multi-artefact distortion meter which can estimate the distortion without prior knowledge of distortion type. We developed the combined distortion meter for full, reduced and no reference approaches to estimate both blockiness and blurriness artefacts. We also studied the impact of spatial masking in image quality estimation. Due to the non linear behaviour of human visual system which perceives different amount of distortions at different spatial frequencies, we applied the masking process which weights the visibility of distortion according to the local spatial activity of the image. We investigated the importance of spatial masking in FR, RR and NR modes. Finally for video sequences, the quality metric of each frame is calculated and then we explored the methods to integrate these scores into a single value. Then the temporal masking is applied to mask the motion by using motion vectors and their standard deviations. The results are compared with the subjective scores provided by LIVE image and video database.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available