An improved motion estimation criterion for temporal coding of video
by Awanish Kumar Mishra; Narendra Kohli
International Journal of Computational Science and Engineering (IJCSE), Vol. 25, No. 3, 2022

Abstract: The size of video data is growing exponentially worldwide and hence there is a need for better video coding standards. MPEG and H.26X have provided several standards for video coding. The latest and effective video coding standards are AVC, HEVC, and AV1. MPEG and H.26X, which employ block matching approaches for temporal coding and the mean absolute difference (MAD) is commonly used as the block matching parameter. MAD is very simple and there is very little complexity in its implementation, but sometimes MAD results in spurious selection of matched block due to different transformations in sequential images and the noise introduced in the frame. To overcome this problem there have been many different matching criteria like vector matching criterion (VMC), smooth constrained mean absolute error (SCMAE) based on DCT, and scaled value criterion (SVC) based on modified pixel values. Criteria described thus far do not take into account the noise generated in the frame, resulting in inaccurate block selection or rejection. A novel motion estimating criterion is proposed in this study and compared to the previous four criteria in terms of PSNR, the average number of assessed search points for every block, and mean MAD for every picture. The suggested matching criterion improves assessed number of search points by about 33.92%, PSNR value by about 9% and average of MAD per pixel by about 70%.

Online publication date: Mon, 30-May-2022

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Computational Science and Engineering (IJCSE):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com