20.500.12556/DKUM-89457 Jedrnat zapis redkih matrik Concise representation of sparse matrices V diplomskem delu opisujemo postopke in implementacijo metod stiskanja redkih matrik. Implementirali smo metode CSR (angl. Compressed Sparse Row), CSF (angl. Coordinate Storage Format), CSV (angl. Compressed Sparse Vector), MSF (angl. Modified Storage Format) in CC (angl. Coordinate Compression). Kot primere redkih matrik smo uporabili decimirane sivinske rastrske slike. Po predstavitvi elementov redke matrike z omenjenimi metodami smo dobljeno zaporedje stisnili z aritmetičnim kodiranjem in z algoritmoma Gzip ter bzip2. Eksperimenti so pokazali, da je metoda CSV najučinkovitejša izmed opisanih metod. In this thesis, procedures and implementation of methods Compressed Sparse Row, Coordinate Storage Format, Compressed Sparse Vector, Modified Storage Format and Coordinate Compression for compressing sparse matrices are described. Decimated grayscale raster images were used as examples of such matrices. The resulting sequences were compressed with arithmetic coding, Gzip, and bzip2 algorithms after representing the sparce matrices by above methods. Experimental results have shown that CSV is the most efficient method. stiskanje podatkov stiskanje koordinat metode stiskanja matrik CSR CSF CSV MSF CC. data compression coordinate compression matrix compression methods Compressed Sparse Row Coordinate Storage Format Compressed Sparse Vector Modified Storage Format Coordinate Compression. true true false [K. Golob] Slovenski jezik Angleški jezik Diplomsko delo/naloga 2024-07-11 07:09:19 2024-09-19 15:26:16 2024-12-23 10:23:34 0000-00-00 00:00:00 2024 0 Maribor 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 004.627(043.2) 220215299 VS_Golob_Klemen_2024.pdf VS_Golob_Klemen_2024.pdf 1 5554725749E93474CCA67FB434FADBFD a9b578c4dcbec37e715bb73f333ab4ec21191451dfe929988767fd7cb99be701 9e0acc26-3f46-11ef-a161-00155d000105 https://dk.um.si/Dokument.php?lang=slv&id=181362 Fakulteta za elektrotehniko, računalništvo in informatiko 0 0 0