The aim of this IWT project, organised by BARCO GRAPHICS, is to develop the software technology for operating a multi-media server for telepublishing applications. The server will be used primarily in pre-press and printing companies where it will provide document management, storage and retrieval functions and image manipulation services to workstations primarily used by document designers. Three main topics can be distinguished in the research activities: database technology, image compression technology and performance issues The contributions of VUB in this project are situated in the domain of image compression: 1. Survey of lossy compression techniques other than JPEG : A broad survey of different lossy compression algorithms will be made, and we will experimentally compare some of the algorithms performance-wise for the different types of image data present in publishing applications. JPEG-based algorithms and segmented image coding are already considered by the other project partners. VUB will consequently consider alternative algorithms which do not belong to either of these two categories, and which have a certain complementarity with respect to the algorithms of the two mentioned categories: VUB will mainly focus on algorithms which avoid the typical drawbacks of JPEG-based algorithms (e.g. inferior image quality of decompressed images due to blocking effects even at relatively low compression ratio), and which are at the same time less computationally demanding than segmented image coding. The practical work of the performance comparisons is quite laborious because of the need to tune the parameters of the compression algorithms for different image-data types. For this reason, a restricted set of algorithms will be selected a priori for performance comparisons A particular case study is the pyramidal multiresolution compression scheme, based on the two-dimensional Discrete Wavelet Transform (2D-DWT). Indeed, an important advantage of the 2D-DWT and other wavelet-based compression techniques is that they do not cause block distortion; also, their computational complexity is comparable to JPEG. A final important advantage of wavelet-based multiresolution compression is that the coding of the image-data is independent of the image size and that it facilitates progressive transmission 2. Peano Scanning of the 3-D color space as an alternative method for color representation : An important aspect of image compression is to develop efficient color representations to improve the compression ratio of the JPEG standard. VUB studies the mapping of three- and four-dimensional color spaces onto one dimension by scanning these spaces using a space-filling curve (Peano Scanning of the 3-D or 4-D color space). This results in an image with scalar pixel values (instead of color-vectors), which can be compressed using JPEG.
|Effective start/end date||1/01/96 → 31/12/98|
Flemish discipline codes
- Mathematical sciences