Title :
Motion compensated in-band prediction for wavelet-based spatially scalable video coding
Author_Institution :
Inst. of Commun. Eng., Aachen Univ., Germany
Abstract :
A wavelet-based spatially scalable video coding scheme with in-band prediction is investigated in combination with JPEG 2000 quantization and arithmetic coding. The wavelet decomposition of each frame of the video data is succeeded by the computation of a blockwise motion compensated prediction for the individual subbands. To achieve half-pixel motion vector accuracy, the shift-invariant overcomplete wavelet transform is exploited. Higher accuracy is obtained by additional interpolation of the overcomplete subbands to maintain spatial scalability. A blockwise decision is made between coding of the motion compensated prediction error (inter mode) or the wavelet coefficients (intra mode). The concept of in-band prediction can be easily combined with a conventional intra-frame wavelet coder. In this approach, JPEG 2000 arithmetic coding is applied to the resulting prediction error.
Keywords :
arithmetic codes; motion compensation; prediction theory; quantisation (signal); transform coding; vectors; video coding; wavelet transforms; JPEG 2000 arithmetic coding; JPEG 2000 quantization; blockwise decision; in-band prediction; motion compensation; motion vector; prediction error coding; spatially scalable video coding; wavelet decomposition; Arithmetic; Band pass filters; Discrete wavelet transforms; Finite impulse response filter; Quantization; Scalability; Streaming media; Transform coding; Video coding; Video compression;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03). 2003 IEEE International Conference on
Print_ISBN :
0-7803-7663-3
DOI :
10.1109/ICASSP.2003.1199110