| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

Show document Help

Title:Primerjava generativnih metod umetne inteligence za kreiranje slikovnih vsebin : magistrsko delo
Authors:ID Ključevšek, Nik (Author)
ID Verber, Domen (Mentor) More about this mentor... New window
Files:.pdf MAG_Kljucevsek_Nik_2025.pdf (21,80 MB)
MD5: 2FA32F3D6C3A53856B7A4DD098865D31
 
Language:Slovenian
Work type:Master's thesis/paper
Typology:2.09 - Master's Thesis
Organization:FERI - Faculty of Electrical Engineering and Computer Science
Abstract:V magistrskem delu smo primerjali štiri priljubljene generativne modele umetne inteligence, ki ustvarjajo slike na podlagi besedilnih opisov: ChatGPT, Stable Diffusion, Midjourney in Microsoft Copilot. Zanimalo nas je, kako se med seboj razlikujejo po kakovosti slik, hitrosti generiranja in zaznavi uporabnikov. Uporabili smo objektivne metrike, izmerili čas generacije in izvedli anketo. Rezultati so pokazali, da Stable Diffusion ustvarja najbolj tehnično natančne slike, Midjourney najbolj prepriča z estetiko, Copilot pa ohranja dobro ravnovesje med obema. ChatGPT kljub nižji tehnični natančnosti pogosto prejme visoke ocene uporabnikov. Ugotovili smo tudi, da se subjektivne ocene ne ujemajo vedno z objektivnimi. Ker se področje umetne inteligence hitro spreminja, bodo nekatere ugotovitve čez čas morda izgubile svojo veljavnost.
Keywords:generativna umetna inteligenca, ustvarjanje slik, primerjava modelov, kakovost slik
Publication status:Published
Publication version:Version of Record
Place of publishing:Maribor
Place of performance:Maribor
Publisher:[N. Ključevšek]
Year of publishing:2025
Number of pages:1 spletni vir (1 datoteka PDF (XIII, 105 str.))
PID:20.500.12556/DKUM-92782 New window
UDC:004.8:004.925.6(043.2)
COBISS.SI-ID:239020035 New window
Publication date in DKUM:06.06.2025
Views:386
Downloads:72
Metadata:XML DC-XML DC-RDF
Categories:KTFMB - FERI
:
Copy citation
  
Average score:(0 votes)
Your score:Voting is allowed only for logged in users.
Share:Bookmark and Share



Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:CC BY-NC-ND 4.0, Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International
Link:http://creativecommons.org/licenses/by-nc-nd/4.0/
Description:The most restrictive Creative Commons license. This only allows people to download and share the work for no commercial gain and for no other purposes.
Licensing start date:13.05.2025

Secondary language

Language:English
Title:Comparison of generative artificial intelligence methods for image content creation
Abstract:In this master’s thesis, we compared four popular generative artificial intelligence models that create images based on text prompts: ChatGPT, Stable Diffusion, Midjourney, and Microsoft Copilot. We focused on how these models differ in terms of image quality, generation speed, and user perception. To evaluate them, we used objective metrics, measured generation times, and conducted a user survey. The results show that Stable Diffusion produces the most technically accurate images, Midjourney stands out for its aesthetics, and Copilot maintains a good balance between the two. Despite lower technical precision, ChatGPT often receives high user ratings. We also found that subjective evaluations do not always align with objective metrics. Since the field of artificial intelligence is evolving rapidly, some findings may become less relevant over time.
Keywords:generative artificial intelligence, image generation, model comparison, image quality


Comments

Leave comment

You must log in to leave a comment.

Comments (0)
0 - 0 / 0
 
There are no comments!

Back
Logos of partners University of Maribor University of Ljubljana University of Primorska University of Nova Gorica