RD-optimized competition scheme for efficient motion prediction

J. Jung; G. Laroche; B. Pesquet

doi:10.1117/12.714374

29 January 2007 RD-optimized competition scheme for efficient motion prediction

J. Jung, G. Laroche, B. Pesquet

Author Affiliations +

Proceedings Volume 6508, Visual Communications and Image Processing 2007; 650815 (2007) https://doi.org/10.1117/12.714374
Event: Electronic Imaging 2007, 2007, San Jose, CA, United States

Abstract

H.264/MPEG4-AVC is the latest video codec provided by the Joint Video Team, gathering ITU-T and ISO/IEC experts. Technically there are no drastic changes compared to its predecessors H.263 and MPEG-4 part 2. It however significantly reduces the bitrate and seems to be progressively adopted by the market. The gain mainly results from the addition of efficient motion compensation tools, variable block sizes, multiple reference frames, 1/4-pel motion accuracy and powerful Skip and Direct modes. A close study of the bits repartition in the bitstream reveals that motion information can represent up to 40% of the total bitstream. As a consequence reduction of motion cost is a priority for future enhancements. This paper proposes a competition-based scheme for the prediction of the motion. It impacts the selection of the motion vectors, based on a modified rate-distortion criterion, for the Inter modes and for the Skip mode. Combined spatial and temporal predictors take benefit of temporal redundancies, where the spatial median usually fails. An average 7% bitrate saving compared to a standard H.264/MPEG4-AVC codec is reported. In addition, on the fly adaptation of the set of predictors is proposed and preliminary results are provided.

Citation Download Citation

J. Jung, G. Laroche, and B. Pesquet "RD-optimized competition scheme for efficient motion prediction", Proc. SPIE 6508, Visual Communications and Image Processing 2007, 650815 (29 January 2007); https://doi.org/10.1117/12.714374

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

;

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE