Stiskanje črno- belih slik z algoritmi delitve prostora

diplomsko delo

Satja Lipuš (Author), Niko Lukač (Mentor)

Abstract

Stiskanje je proces zmanjševanja podatkov v pomnilniškem prostoru. V diplomskem delu predstavimo metodo stiskanja črno-belih slik z uporabo algoritmov, ki delijo prostor. Omejili smo se na sledeče tri metode deljenja prostora: enakomerna mreža, štiriško drevo in Kd drevo. Enakomerna mreža temelji na matematičnem modelu enakomerne kartezične mreže. Medtem, ko štiriško drevo rekurzivno deli prostor v štiri podprostore v obliki štiriškega podatkovnega drevesa, katerega vozlišča prestavljajo nadaljnjo delitev v globino. Kd drevo deli prostor v binarno drevo glede na mediano vrednosti, pri čemer vsako nadaljno rekuzijo izmenjuje dimenzijo delitve. Imenovani algoritmi predstavljajo obliko izgubnega stiskanja z različno učinkovitostjo. V diplomskem delu smo primerjali izgubo podatkov pri stiskanju poljubnih slik s predstavljenimi algoritmi in le to kvantitativno ocenili z uporabo metode povrečne kvadratne napake in strukturne podobnosti. Pri rezultatih smo ugotovili, da je pri delitvi prostora optimalna izbira štiriško drevo.

Keywords

stiskanje;razširjanje;črno-bele slike;delitev prostora;podatkovna drevesa;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2019
Typology:	2.11 - Undergraduate Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[S. Lipuš]
UDC:	004.932.021(043.2)
COBISS:	22927638
Views:	722
Downloads:	37
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Compression of binary images by using space partitioning algorithms
Secondary abstract:	Compression is a process of minimizing needed data to store given information on the computer. The thesis addresses one of the known methods of compression named space-division and applies it to a series of black and white images. We restricted ourselves to three space-division methods, namely Uniform Grid, Quad Tree and Kd Tree. Uniform Grid is based on the mathematical model of the Uniform Cartesian Grid. The method divides the space recursively into a uniform grid, independent of the values inside the space. Meanwhile the Quad Tree divides space in a form of a Quad Data Tree, where its nodes are further divisions while the leafs represent the final divided space. Kd Tree divides space based on the median of the spaces dimensions interchangeably, storing its data similarly to Quad Tree into a Binary Tree. The named algorithms represent a form of lossy compression with differing effectiveness. We compared their data loss compressing given images and described it with the structural similarity method. Looking at the results we concluded that that in regards to spacial division algorithm, the optimal choice is the quad tree.
Secondary keywords:	compression;decompression;monochromatic image;space-division;data trees.;
Type (COBISS):	Bachelor thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages:	V, 31 f.
ID:	11214852