Hurtgen,B., Stiller,C. : 'Fast hierarchical codebook search for fractal coding of still images'. EOS/SPIE Visual Communication and PACS for Medical Applications 93, 1993.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....about 325x10 9 floating point operations (flops) such a very huge number of flops clearly explains the complexity of the exact solution of the IFS inverse problem. In order to reduce the number of computations required, some sub optimal techniques have been proposed. For example, in [12] and [5] the search is not performed over the whole space (the domain pool, DP) but in a small subset of DP (determined through a neighbourhood basis) An alternative way to reduce DP is classification [7] 8] in such a the case dimensions of DP are reduced by considering only elements belonging to a ....

Hurtgen,B., Stiller,C. : 'Fast hierarchical codebook search for fractal coding of still images'. EOS/SPIE Visual Communication and PACS for Medical Applications 93, 1993.

....tree for nearest neighbors of each one of them. The preprocessing time to create the data structure for the multi dimensional search is not a limitation of the method as demonstrated by our experiments. A forerunner of feature vectors as described above has been presented by Hurtgen and Stiller [HuSt93]. As in the classification of Fisher, Jacobs, and Boss an image block is partitioned into its four quadrants and their mean intensities are computed. Then a vector consisting of four bits is constructed as follows: the i th bit is 1 if Fractal Image Compression An Introductory Overview 27 ....

.... Overview 35 For example, in the work that followed Monro and Dudbridge [MoDu92a] for each range the codebook block to be used to cover the range is uniquely predetermined to be a specific block that contains the range block [WoMo95] A similar idea has been pursued by Hurtgen and Stiller [HuSt93] where the search area for a domain is restricted to a neighborhood of the current range. Additionally, a few sparsely spaced domains far from the range are taken into account as an option. Iterated Systems, Inc. seems also to prefer a local searching [GeLu96] In [Saup96a] we considered a ....

Hurtgen, B., Stiller, C., Fast hierarchical codebook search for fractal coding of still images, in: EOS/SPIE Visual Communications and PACS for Medical Applications'93, Berlin, 1993.

....image blocks in Lenna. In the figure, dark areas are smooth , gray areas are moderate and bright areas rugged . They represent 27.34 , 42.19 and 30.47 of the image, respectively. Fig. 1: Classification of image blocks In our implementation, we employed Hurtgen s spiral shaped search area [5] to carry out a limited search for the best domain blocks. We use 7, 5, 2, and 5 bits to uniformly quantize DC terms, indices for best domain blocks, isometries, and scaling factors, respectively. Figures 2 5 show distortion (measured as PSNR) as a function of image compression for the whole image ....

Hurten, B. and Stiller C., Fast Hierarchical Codebook Search for Fractal Coding of Still Images, SPIE Visual Communications and PACS for Medical Applications'93, Berlin, Germany, 1993

....the interconnection network) of the target architecture. 5 Future research In a subsequent paper we investigate the influence of the use of sequential speedup methods (like quadrant brightness and quadrant variance classification schemes [8] clustering techniques [23] geometric search techniques [3, 15] and other advanced search techniques) on the parallelization techniques introduced in this paper. Whereas algorithms A can easily cope with structured domain pools [13] or different search techniques (at least in the search procedure itself) the combination of such methods and parallelization ....

B. Hurtgen and C. Stiller. Fast hierarchical codebook search for fractal coding of still images. In SPIE Visual Communication and PACS for Medical Applications, pages 397--408, Berlin, Germany, 1993.

....blocks requires about 325x10 9 floating point operations(flops) such a very huge number of flops clearly explains the complexity of the exact solution of the IFS inverse problem. In order to reduce the amount of computations required, some sub optimal techniques were proposed. In [Mon 92a] and [Hur 93] for example, the search is not performed over the whole space (the domain pool, DP) but in a small subset of DP (determined through a neighbourhood basis) An alternative way to reduce DP is classification [Jaq 93] Jac 92] in such a case dimensions of DP are reduced by introducing in it only ....

Hurtgen,B., Stiller,C. : 'Fast hierarchical codebook search for fractal coding of still images'. EOS/SPIE Visual Communication and PACS for Medical Applications 93, 1993.

....subset of all possible domains. For example, in the work that followed Monro and Dudbridge [28] for each range the codebook block to be used to cover the range is uniquely predetermined to be a specific block that contains the range block. A similar idea has been pursued by Hurtgen and Stiller [19] where the search area for a domain is restricted to a neighborhood of the current range or additionally a few sparsely spaced domains far from the range are taken into account as an option. In [35] we have noticed that one can discard domain blocks with low variance values with a little ....

....straight usage of the projected and normalized codebook blocks as in this paper, the attempt of using invariant moments of ranges and domains, and tree structured methods. 7. 1 Straight feature vectors A forerunner of feature vectors as used in this paper has been presented by Hurtgen and Stiller [19]. As in the classification of Fisher, Jacobs, and Boss an image block is partitioned into its four quadrants and their mean intensities are computed. Then a vector consisting of four bits is constructed as follows: the i th bit is 1 if the mean of the i th quadrant is above the overall mean, and 0 ....

[Article contains additional citation context not shown here]

Hurtgen, B., Stiller, C., Fast hierarchical codebook search for fractal coding of still images, EOS/SPIE Visual Communications and PACS for Medical Applications '93, Berlin (1993) 397--408.

No context found.

H¨urtgen, B., Stiller, C., Fast hierarchical codebook search for fractal coding of still images, in: EOS/SPIE Visual Communications and PACS for Medical Applications'93, Berlin, 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC