diplomsko delo
Gregor Babnik (Author), Luka Šajn (Mentor)

Abstract

Delo obravnava način samodejnega štetja objektov na slikah. Uporabljena metoda za učenje je genetski algoritem, s katerim se išče zaporedje ustreznih operacij, ki se jih nato izvede nad podanimi slikami. Uspešnost posamezne rešitve se meri z odstopanjem med številoma preštetih in dejanskih objektov. Za nastavitev ločljivosti vhodnih slik se uporablja algoritem ARes. Implementacija procesiranja slik se izvaja z uporabo programskih knjižnic Tensorflow in OpenCV. Delo je testirano na množicah slik iz različnih domen.

Keywords

štetje;genetski algoritem;tensorflow;opencv;ares;računalništvo in informatika;univerzitetni študij;diplomske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.11 - Undergraduate Thesis
Organization: UL FRI - Faculty of Computer and Information Science
Publisher: [G. Babnik]
UDC: 004(043.2)
COBISS: 21798403 Link will open in a new window
Views: 863
Downloads: 157
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Counting objects in images using a genetic algorithm
Secondary abstract: The work deals with the automatic counting of objects in images. A genetic algorithm is used as a learning method to find appropriate operations used to process the images. The success of an individual solution is measured as a difference between the number of counted objects and the real object count. ARes algorithm is used to adjust the resolution of input images. The image processing part is implemented using two libraries TensorFlow and OpenCV. The work is tested against various sets of images in different domains.
Secondary keywords: counting;genetic algorithm;tensorflow;opencv;ares;computer and information science;diploma;
Type (COBISS): Bachelor thesis/paper
Study programme: 1000468
Embargo end date (OpenAIRE): 1970-01-01
Thesis comment: Univ. v Ljubljani, Fak. za računalništvo in informatiko
Pages: 72 str.
ID: 11880848