Izboljšava kontrasta slik na grafičnih procesnih enotah

diplomsko delo

Gregor Černak (Author), Patricio Bulić (Mentor)

Abstract

Izboljšava kontrasta slik je pomembna pri slikah, ki so bile posnete pri nizki svetlobi. Najbolj uporabljena metoda v procesiranju slik je izenačenje histograma. V diplomskem delu bomo predstavili sekvenčno in vzporedno implementacijo dveh tehnik. Ti tehniki sta prilagodljivo izenačenje histograma in kontrastno omejeno prilagodljivo izenačenje histograma. Za obe tehniki bomo zasnovali algoritem za sekvenčno in vzporedno implementacijo. Sekvenčno implementacijo bomo izvajali na centralni procesni enoti, medtem ko bomo vzporedno implementacijo s pomočjo ogrodja OpenCL, ki omogoča standardni vmesnik za vzporedno procesiranje, izvajali na grafični procesni enoti. Implementaciji posameznih algoritmov bomo primerjali na različni strojni opremi. Pridobljene rezultate bomo predstavili in opisali lastnosti posamezne implementacije, ter prišli do ugotovitve, pri kateri meji je časovno bolj obetavna vzporedna implementacija algoritma.

Keywords

adaptive histogram equalization;contrast limited adaptive histogram equalization;OpenCL;vzporedna implementacija;GPU;računalništvo;računalništvo in informatika;visokošolski strokovni študij;diplomske naloge;

Data

Language:	Slovenian
Year of publishing:	2021
Typology:	2.11 - Undergraduate Thesis
Organization:	UL FRI - Faculty of Computer and Information Science
Publisher:	[G. Černak]
UDC:	004(043.2)
COBISS:	54174723
Views:	321
Downloads:	192
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	English
Secondary title:	Contrast enhancement of images on general-purpose graphic processing units
Secondary abstract:	Improving the contrast is important for images that are taken in low light. The most common method used in image processing is histogram equalization. In the thesis work, we will present sequential and parallel implementation of both techniques. These techniques are adaptive histogram equalization and contrast-limited adaptive histogram equalization. For both techniques, we describe the algorithm. The sequential implementation is simple and runs on the central processing unit. Parallel implementation uses the OpenCL framework, which is typically used in parallel processing interfaces and runs our code on the graphics processing unit. We compare the implementaitons of individual algorithms, present the results and describe the specifics. Finally, we reach the conclusion of when to use each implementation and at which point the parallel implementation of the algorithm is executed faster.
Secondary keywords:	Adaptive histogram equalization;Contrast limited adaptive histogram equalization;OpenCL;parallel implementation;GPU;
Type (COBISS):	Bachelor thesis/paper
Study programme:	1000470
Embargo end date (OpenAIRE):	1970-01-01
Thesis comment:	Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages:	68 str.
ID:	12615434