Statistical analysis of the 2D-DCT coefficients of the differential signal for images

2010 Data Compression Conference

2010

Current video coding schemes employ motion compensation to exploit the fact that the signal forms an auto-regressive process along the motion trajectory, and remove temporal redundancies with prior reconstructed samples via prediction. However, the decoder may, in principle, also exploit correlations with received encoding information of future frames. In contrast to current decoders that reconstruct every block immediately as the corresponding quantization indices are available, we propose an estimation-theoretic delayed decoding scheme which leverages quantization and motion information of one or more future frames to refine the reconstruction of the current block. The scheme, implemented in the transform domain, efficiently combines all available (including future) information in an appropriately derived conditional pdf, to obtain the optimal delayed reconstruction of each transform coefficient in the frame. Experiments demonstrate substantial gains over the standard H.264 decoder. The scheme learns the autoregressive model from information available to the decoder, and compatibility with the standard syntax and existing encoders is retained.

Section: Statistical Modelmentioning

confidence: 99%

Estimation-Theoretic Delayed Decoding of Predictively Encoded Video Sequences

2010 Data Compression Conference

2010

“…The innovation probability density function (pdf) denoted by pZ (z) is assumed to be Laplacian in keeping with prior work [6], i.e.,…”

Section: Optimal Enhancement Layer Prediction In Svcmentioning

confidence: 99%

“…In the above equation the zero-delay pdf of xn, p(xn|x b n−1 , I b n ), is weighed by the probability p(I b n+1 |xn) of the known future outcome to obtain the 1-sample delayed pdf on the LHS of (6), that incorporates all known information at the decoder up to a delay of 1 sample. Now,…”

Section: Delayed Decoding Of Predictively Encoded Videomentioning

confidence: 99%

Estimation-theoretic approach to delayed prediction in scalable video coding

2010 IEEE International Conference on Image Processing

2010

Scalable video coding (SVC) employs inter-frame prediction at the base and/or the enhancement layers. Since the base layer can be encoded/decoded independent of the enhancement layers, we consider here the potential gains when prediction at the enhancement layers is delayed to accumulate and incorporate additional future information from the base layer. We build on two basic estimationtheoretic (ET) approaches developed by our group: an ET approach for enhancement layer prediction that optimally combines current base layer with prior enhancement layer information, and our recent ET approach for delayed decoding. The proposed technique fully exploits all the available information from the base layer, including any future frame information, and past enhancement layer information. It achieves considerable gains over zero-delay techniques including both standard SVC, and SVC with optimal ET prediction (but with zero encoding delay).

“…As an aside, it should be noted that the zero-mean innovations in (1) imply that x n is itself zero-mean, whenever |ρ| < 1 (i.e., any non-zero means during initialization of the process are eventually damped down by ρ). It was indeed observed in the experiment that the mean of DCT coefficients at any AC frequency was always nearly zero.…”

Section: Transform Domain Prediction: Model and Motivationmentioning

confidence: 99%

“…In both [4] and [10], such a viewpoint was necessary to explicitly account for the quantization interval information exploited by the ET framework, which is available only in the transform domain. The innovation of each scalar AR process (per frequency) was modeled as Laplacian [1], and the temporal correlation coefficient of each AR process at different spatial frequencies was assumed to be unity (as is the common practice in pixel domain). Note that in this case the transform domain model is congruent with the pixel domain AR model due to the unitarity of the spatial transform.…”

Section: Introductionmentioning

confidence: 99%

Transform-domain temporal prediction in video coding: Exploiting correlation variation across coefficients

2010 IEEE International Conference on Image Processing

2010

Abstract-Temporal prediction in standard video coding is performed in the spatial domain, where each pixel block is predicted from a motion-compensated pixel block in a previously reconstructed frame. Such prediction treats each pixel independently and ignores underlying spatial correlations. In contrast, this paper proposes a paradigm for motion-compensated prediction in the transform domain, that eliminates much of the spatial correlation before individual frequency components along a motion trajectory are independently predicted. The proposed scheme exploits the true temporal correlations, that emerge only after signal decomposition, and vary considerably from low to high frequency. The scheme spatially and temporally adapts to the evolving source statistics via a recursive procedure to obtain the cross-correlation between transform coefficients on the same motion trajectory. This recursion involves already reconstructed data and precludes the need for any additional side-information in the bit-stream. Experiments demonstrate substantial performance gains in comparison with the standard codec that employs conventional pixel domain motion-compensated prediction.