prikaz prve stranice dokumenta Analiza i usporedba kvalitete 3D modela generiranih AI alatima
  Download
PDF 2.88 MB
master's thesis
Analiza i usporedba kvalitete 3D modela generiranih AI alatima
Koprivnica: University North, 2024. urn:nbn:hr:122:501874
Vusić, Bernard
University North
University centre Varaždin
Department of Multimedia, Design and Application

Institutional repository: University North Digital Repository

Cite this document

Vusić, B. (2024). Analiza i usporedba kvalitete 3D modela generiranih AI alatima (Master's thesis). Koprivnica: University North. Retrieved from https://urn.nsk.hr/urn:nbn:hr:122:501874

Vusić, Bernard. "Analiza i usporedba kvalitete 3D modela generiranih AI alatima." Master's thesis, University North, 2024. https://urn.nsk.hr/urn:nbn:hr:122:501874

Vusić, Bernard. "Analiza i usporedba kvalitete 3D modela generiranih AI alatima." Master's thesis, University North, 2024. https://urn.nsk.hr/urn:nbn:hr:122:501874

Vusić, B. (2024). 'Analiza i usporedba kvalitete 3D modela generiranih AI alatima', Master's thesis, University North, accessed 15 November 2024, https://urn.nsk.hr/urn:nbn:hr:122:501874

Vusić B. Analiza i usporedba kvalitete 3D modela generiranih AI alatima [Master's thesis]. Koprivnica: University North; 2024 [cited 2024 November 15] Available at: https://urn.nsk.hr/urn:nbn:hr:122:501874

B. Vusić, "Analiza i usporedba kvalitete 3D modela generiranih AI alatima", Master's thesis, University North, Koprivnica, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:122:501874

Metadata
TitleAnaliza i usporedba kvalitete 3D modela generiranih AI alatima
Title (english)Analysis and comparison of the quality of 3D models generated by AI tools
AuthorBernard Vusić
MentorAndrija Bernik (mentor)
Committee memberDean Valdec (predsjednik povjerenstva)
Committee memberAndrija Bernik (član povjerenstva)
Committee memberEmil Dumić (član povjerenstva)
GranterUniversity North
(University centre Varaždin)
(Department of Multimedia, Design and Application)
Koprivnica
Defense date and country2024-10-29, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Graphic Technology
Processes of Graphic Reproduction
Abstract
Tema diplomskog rada fokusira se na analizu i usporedbu performansi različitih AI alata za generiranje 3D modela, s naglaskom na njihovu sposobnost preciznog odgovora na zadane specifikacije (promptove) i zahtjeve industrije. Uz primjenu naprednih algoritama kao što su Neural Radiance Fields (NeRF) i Generative Adversarial Networks (GAN), alati poput DreamFusion, Magic3D i 3DFY.ai omogućuju automatizirano stvaranje 3D modela s visokim stupnjem detalja i realizma. U radu se ocjenjuju komercijalni alati, uključujući Meshy, Studio Masterpiece, Luma AI, Tripo AI, 3D AI Studio, NVIDIA Edify-3D i Sloyd, kroz kriterije kao što su geometrijska točnost, kvaliteta tekstura, proporcionalnost i ukupna vjernost prema promptu. Rad istražuje koliko su navedeni alati sposobni zadovoljiti složene zahtjeve za preciznim modeliranjem, ocjenjujući ih sustavom bodovanja koji rangira svaki alat u odnosu na ključne parametre kvalitete. Posebna pažnja posvećena je ocjeni kvalitete tekstura, topologiji modela i njegovoj geometrijskoj točnosti kako bi se utvrdilo koji alati pružaju najbolje rezultate za specifične industrijske primjene.
Abstract (english)
The topic of this thesis focuses on analyzing and comparing the performance of various AI tools for 3D model generation, emphasizing their ability to precisely respond to specified prompts and industry requirements. Through the use of advanced algorithms such as Neural Radiance Fields (NeRF) and Generative Adversarial Networks (GAN), tools like DreamFusion, Magic3D, and 3DFY.ai enable the automated creation of 3D models with a high level of detail and realism. The thesis evaluates commercial tools, including Meshy, Studio Masterpiece, Luma AI, Tripo AI, 3D AI Studio, NVIDIA Edify-3D, and Sloyd, based on criteria such as geometric accuracy, texture quality, proportionality, and overall fidelity to the prompt. The study explores how well these tools meet complex demands for precise modeling, employing a scoring system that ranks each tool against key quality parameters. Special attention is given to assessing texture quality, model topology, and geometric accuracy to determine which tools deliver the best results for specific industrial applications.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:122:501874
Study programmeTitle: Multimedia Science
Study programme type: university
Study level: graduate
Academic / professional title: sveučilišni magistar/magistra inženjer/inženjerka multimedijske i grafičke tehnologije (sveučilišni magistar/magistra inženjer/inženjerka multimedijske i grafičke tehnologije)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2024-11-13 11:32:00