prikaz prve stranice dokumenta Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže
  Download
PDF 1016.61 KB
undergraduate thesis
Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže
2020. urn:nbn:hr:200:415614
Stipić, Nikola
Josip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
Department of Computer Engineering and Automation
Chair of Automation and Robotics

Cite this document

Stipić, N. (2020). Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže (Undergraduate thesis). Retrieved from https://urn.nsk.hr/urn:nbn:hr:200:415614

Stipić, Nikola. "Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. https://urn.nsk.hr/urn:nbn:hr:200:415614

Stipić, Nikola. "Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže." Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, 2020. https://urn.nsk.hr/urn:nbn:hr:200:415614

Stipić, N. (2020). 'Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže', Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, accessed 20 March 2023, https://urn.nsk.hr/urn:nbn:hr:200:415614

Stipić N. Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže [Undergraduate thesis]. Osijek: Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek; 2020 [cited 2023 March 20] Available at: https://urn.nsk.hr/urn:nbn:hr:200:415614

N. Stipić, "Semantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže", Undergraduate thesis, Josip Juraj Strossmayer University of Osijek, Faculty of Electrical Engineering, Computer Science and Information Technology Osijek, Osijek, 2020. Available at: https://urn.nsk.hr/urn:nbn:hr:200:415614

Please login to the repository to save this object to your list.
Metadata
TitleSemantička segmentacija 3D snimki unutrašnjih prostora primjenom umjetne neuronske mreže
Title (english)Semantic Segmentation of 3D Scans of Indoor Environments Using Artificial Neural Networks
AuthorNikola Stipić
MentorRobert Cupec (mentor)
Committee memberRobert Cupec (predsjednik povjerenstva)
Committee memberZdravko Krpić (član povjerenstva)
Committee memberEmmanuel Karlo Nyarko (član povjerenstva)
GranterJosip Juraj Strossmayer University of Osijek
Faculty of Electrical Engineering, Computer Science and Information Technology Osijek
(Department of Computer Engineering and Automation)
(Chair of Automation and Robotics)
Osijek
Defense date and country2020-09-24, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Artificial Intelligence
Abstract
U ovom radu istražena je i implementirana jedna od metoda za semantičku segmentaciju 3D snimaka unutrašnjih prostora umjetnom neuronskom mrežom. Opisane su neke od prethodno korištenih te „State of the Art“ metode za ovaj problem. Detaljno je proučena arhitektura R-PointNet neuronske mreže te svi procesi korišteni u Generative Shape Proposal Network (GSPN) mreži, koja predstavlja sastavni dio R-PointNet mreže. Navedeni su i ukratko opisani neki od najpoznatijih 3D ispitnih podatkovnih skupova unutrašnjih prostorija i njihova primjena za učenje i ispitivanje GSPN metode. Provedena je eksperimentalna evaluacija s učenjem neuronske mreže na osobnom računalu te analizirani rezultati uspješnosti. Istaknute su mogućnosti i nedostatci semantičke segmentacije u unutrašnjim prostorima te njihova eventualna primjena.
Abstract (english)
This paper investigates and implements one of the methods for semantic segmentation of 3D in-door scenes using artificial neural network. Some State of the Art solutions for this problem are described. Architecture of the neural network R-PointNet and all of the processes used in Generative Shape Proposal Network (GSPN) network, representing the main part of the R-PointNet, are studied in detail. Some of the most popular datasets of 3D in-door scenes are listed and briefly explained as well as their use for training and testing of the GSPN method. An experimental evaluation was performed where the considered neural network was trained on a personal computer and the resulting efficiency scores were analyzed. Some advantages and disadvantages of the in-door semantic segmentation using the considered neural network are discussed and its possible real-world applications are suggested.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:200:415614
Study programmeTitle: Undergraduate University Study Programme in Computer Engineering
Study programme type: university
Study level: undergraduate
Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2020-11-13 14:01:32