pace by narrowing the scope of the research and enabling seamless integration of this work into inherently adaptive filtering process. Some of the benefits of this modular and adaptive framework have been presented on the example of the MPEG-4 postprocessing improvements.

by block discrete cosine transform (DCT) methods (e.g., JPEG, MPEG) suffer from blocking and ringing. Postprocessing to reduce these artifacts has appeal since it only requires the decompressed image. Proposed methods include block-boundary filtering [1], convex projections (CP) [2, 3], and maximum a

artifacts in the corrected sequence result otherwise. We show that the performance of blotch detectors can be improved significantly by taking statistical influence of noise on the detection mechanism into account. Further improvements are achieved first by using a double-stage detection strategy and second

Despite significant improvements in image quality over the past several years, the full exploitation of magnetic resonance image (MRI) data is often limited by low signal to noise ratio (SNR) or contrast to noise ratio (CNR). In implementing new MR techniques, the criteria of acquisition speed

combines hardware instrumentation to measure current level with kernel software support to perform statistical sampling of system activity. Postprocessing software maps the sample data to program structure and produces a profile of energy usage by process and procedure. Using PowerScope, we have been able

of the enhancement methods for speaker verification, and compare them against a Gaussian target mapping implementation. Results indicate improvements of the warping technique over a number of methods such as Cepstral Mean Subtraction (CMS), modulation spectrum processing, and short-term windowed CMS and variance

such that the deblocking to be performed can be taken into account in the compression and the two processes can be jointly optimised. An example is also provided to show how this idea works successfully in a MPEG-4 codec to boost the ratedistortion performance when the suggested deblocking filter is exploited in the post-processing

such that the statistical behavior is improved. For video object (VO) shape coding, we use an adaptive threshold method to remove shape-coding artifacts for MPEG-4 applications. A dynamic bit allocation among VOs is implemented according to the coding complexities for each VO. SRC achieves more accurate bit allocation

such that the deblocking to be performed can be taken into account in the compression and the two processes can be jointly optimised. An example is also provided to show how this idea works successfully in a MPEG-4 codec to boost the rate-distortion performance when the suggested deblocking filter is exploited in the post-processing

coding or post-processing can reveal originally masked distortions. A new emerging standard MPEG-4 ALS introduces an efficient lossless audio coding that enables the compression of digital audio data without any loss in quality and the perfect reconstruction of the original signal. MPEG-4 ALS employs