master's thesis
Image retrieval system based on simplified image histograms

Adrian Bece (2015)
Metadata
TitleSustav pretraživanja slika korištenjem histograma pojednostavljene slike
AuthorAdrian Bece
Mentor(s)Goran Martinović (thesis advisor)
Abstract
U ovom diplomskom radu je opisan i programski ostvaren algoritam za dohvaćanje slike iz baze podataka temeljen na usporedbi histograma slike. Opisani su, predloženi i ugrađeni algoritmi formiranja histograma i usporedbe histograma, te je prikazana struktura baze podataka i način pohrane slika u njoj. Prikazani su rezultati ispitivanja, kao i prednosti i nedostatci svake pojedine metode i konačnog algoritma. Neke od prednosti su brzina izvođenja algoritma ako se koristi paralelizam i pohrana slike u priručnoj memoriji, te robusnost algoritama na rotaciju slike, zrcaljenje, promjenu razlučivosti slike i zamućenje. Uočeni nedostatci algoritma su smanjena robusnost na promjenu vrijednosti piksela (osvjetljenje ili kontrast), zanemarivanje informacije o teksturama na slici, te postojanje nekih posebnih slučajeva pogrešnog dohvaćanja slike.
Keywordsimage processing algorithm image analysis global descriptor global histogram image histogram
Parallel title (English)Image retrieval system based on simplified image histograms
Committee MembersGoran Martinović (committee member)
Alfonzo Baumgartner (committee member)
Emmanuel Karlo Nyarko (committee member)
GranterSveučilište Josipa Jurja Strossmayera u Osijeku
Fakultet elektrotehnike, računarstva i informacijskih tehnologija Osijek
Lower level organizational unitsZavod za programsko inženjerstvo
Katedra za programske jezike i sustave
PlaceOsijek
StateCroatia
Scientific field, discipline, subdisciplineTECHNICAL SCIENCES
Computing
Architecture of Computer Systems
Study programme typeuniversity
Study levelgraduate
Study programmeGraduate University Study Programme in Computer Engineering
Academic title abbreviationmag.ing.comp.
Genremaster's thesis
Language Croatian
Defense date2015-09-15
Parallel abstract (English)
This thesis describes an image retrieval algorithm based on the comparison of the image histograms accompanied by the description of the application based on the mentioned algorithm. The algorithms forming the histogram are described and implemented. Additionally, the histogram comparison algorithms, database structure and methods of storing images into a database are provided. The thesis shows performance test results, as well as the advantages and disadvantages of every conversion method and the algorithm itself. Some of the advantages are the speed of execution when using a parallelism and cache memory on the one and robustness to image rotation, mirroring, resolution change and blurring on the other hand. In comparison, some of the disadvantages include decreased robustness to a pixel value change (brightness and contrast), ignoring information about textures and image and some specific cases of incorrect image retrieval.
Parallel keywords (Croatian)algoritmi obrade slike analiza slike globalni deskriptor globalni histogram histogram slike.
Resource typetext
Access conditionOpen access
Terms of usehttp://rightsstatements.org/vocab/InC/1.0/
URN:NBNhttps://urn.nsk.hr/urn:nbn:hr:200:275375
CommitterAnka Ovničević