1 October 2004 Low bit rate wavelet video codec using edge based motion estimation and context-based coding
Author Affiliations +
J. of Electronic Imaging, 13(4), (2004). doi:10.1117/1.1786296
Abstract
This paper presents a new wavelet-based video codec for very low bit rate coding, where each frame of the test sequence is transformed using discrete wavelet transformation. The hierarchical edge based motion estimation is applied between adjacent transformed frames. The aim is to reduce the computational complexity of a motion estimation algorithm and to preserve coding efficiency. The edge detection is incorporated into the motion estimation process in order to determine the moving regions. The edge is assumed as a point between two different moving areas in the proposed motion estimation algorithm. All transformed frames are quantized and prediction errors are coded using quantization and probability models designed with fuzzy logic. This proposed method is comparable with the H.26L video coding standard in the rate distortion sense, and has lower computational complexity than the H.26L standard.
Dušan Gleich, Peter Planinšic, Zarko Cucej, "Low bit rate wavelet video codec using edge based motion estimation and context-based coding," Journal of Electronic Imaging 13(4), (1 October 2004). https://doi.org/10.1117/1.1786296
JOURNAL ARTICLE
11 PAGES


SHARE
RELATED CONTENT

Macro motion vector quantization
Proceedings of SPIE (April 21 1995)
Interframe vector wavelet coding technique
Proceedings of SPIE (January 10 1997)

Back to Top