A Novel Adaptive Frame Determination Methodology with Fixed and Adaptive Thresholds Using Affine Motion Parameter for Video Quality Enhancement

ABSTRACT

There is shortage of novel methodologies in obtaining a better quality compressed video. Diverse exploration resulted in Group of pictures aiming at Intra, Bi-directional & Predicted frame identification for performing compression. The conventional GoP practices fixed frame pattern irrespective of the movement by the camera. This paper deals with the choice of adaptive frame identification by incorporating fixed and empirical values. The fixed values are chosen based on the various motion types. NSEW affine translation is incorporated to substitute B frames obtained from fixed and empirical threshold values with I or P frame. Motion Estimation & warping using affine are preferred for compression and decompression to obtain the compressed data and the warped. Four metrics explicitly have been analyzed to demonstrate the superiority of the suggested procedure.

KEYWORDS:

• Adaptive frame determination
• B frames
• compression
• File size
• Fixed Group of pictures
• Motion estimation using affine
• Structural similarity index
• Warping using affine

Acknowledgment

Authors express their sincere appreciation to Karunya Institute of Technology and Sciences for the research ambience and infrastructure that has been extended to them.

Additional information

Notes on contributors

D. Raveena Judie Dolly

D Raveena Judie Dolly received her BE degree from Karunya Institute of Technology, Coimbatore in 2004 and her ME degree from Karunya Deemed University in 2006. She has been working since them in various engineering colleges. Her research interest include image processing, communication systems and medical imaging.

G. Josemin Bala

G Josemin Bala completed her doctorate in Anna University and is currently working as a professor in the Dept of Electronics & Communication Engineering, Karunya University. Her research interests include wireless communication, VLSI design and communication. She also serves as a reviewer for number of peerreviewed journals. Email: [email protected]

J. Dinesh Peter

J Dinesh Peter completed his doctorate in NIT, Calicut and is currently working as an associate professor in theDept of Information Technology, Karunya University. His research interests include image processing, medical imaging He also serves as a reviewer for number of peer-reviewed journals. Email: [email protected]

D. J. Jagannath

D J Jagannath from the Queen of Hills – The Nilgiris. He graduated in engineering, specializing in electronics & communication under the University of Madras during 1998. He then pursued masters and doctorate in engineering, specializing in applied electronics at Karunya University. His research is in the field of biomedical signal processing amalgamated with computational intelligence.He has publishedhisworks in several international journals with an impact factor range of 1.4–2.810 He is now serving at Karunya University in the Department of Electronics and Communication Engineering as an assistant professor. Email: [email protected]