Search academic texts

Information × Registration Number 0211U006584, 0110U004001 , R & D reports Title Improving image compression loss less in modern graphics formats popup.stage_title Head Bomba Andriy Yaroslavovych, Registration Date 11-05-2011 Organization Rivne State Humanitarian University popup.description2 A research object is processes of compression of data without losses. A research purpose is an increase of efficiency of compression of images in the format of PNG by perfection and account of co-operation of used and application of alternative or new methods and code algorithms.The thesis is to devoted the question of rising the efficiency of compression without losses in the graphic formats (the PNG format as an example) of rasters, which use predictors, dictionary algorithm LZ77, context-independent coding and their combinations by means of improving and taking into consideration of the interaction of them and application of alternative or new methods and code algorithms. The method of searching of consilient sequences of the flow of data by choosing the shortest hash-chains and the method of exact computation of sizes of blocks of dynamic codes of HUFF has been improved for accelerating of coding in the PNG format. For diminishing of coefficient of compression in the ratified standard PNG format on the average of more than 3.9 % for the first time the postprocess method of reduction of size of the compressed block in the format of DEFLATE and method of previous analysis of images by breaking up into minimum and homogeneous blocks of lines have been realized. As for bringing of modifications in this format methods for the generation of the differential coloured models with the whole as well as real coefficients have been elaborated. The structure of curriculum of algorithm of LZ77 has been improved and a palette for group statistical coding which enabled to decrease coefficient of compression for more than 12 % in the case of compatible realization with the algorithm of correction of values of predictors. Keywords: color image, compression without losses, graphic PNG format, methods of the previous processing of images. Product Description popup.authors Бомба Андрій Ярославович Шпортько Леся Василівна Шпортько Олександр Володимирович popup.nrat_date 2020-04-02 Close