prikaz prve stranice dokumenta Konvolucijski modeli za panoptičku segmentaciju
No public access
undergraduate thesis
Konvolucijski modeli za panoptičku segmentaciju
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2022. urn:nbn:hr:168:243394
Pavković, Tomislav
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Pavković, T. (2022). Konvolucijski modeli za panoptičku segmentaciju (Undergraduate thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:243394

Pavković, Tomislav. "Konvolucijski modeli za panoptičku segmentaciju." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2022. https://urn.nsk.hr/urn:nbn:hr:168:243394

Pavković, Tomislav. "Konvolucijski modeli za panoptičku segmentaciju." Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2022. https://urn.nsk.hr/urn:nbn:hr:168:243394

Pavković, T. (2022). 'Konvolucijski modeli za panoptičku segmentaciju', Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 23 July 2024, https://urn.nsk.hr/urn:nbn:hr:168:243394

Pavković T. Konvolucijski modeli za panoptičku segmentaciju [Undergraduate thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2022 [cited 2024 July 23] Available at: https://urn.nsk.hr/urn:nbn:hr:168:243394

T. Pavković, "Konvolucijski modeli za panoptičku segmentaciju", Undergraduate thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2022. Available at: https://urn.nsk.hr/urn:nbn:hr:168:243394

Metadata
TitleKonvolucijski modeli za panoptičku segmentaciju
Title (english)Convolutional models for panoptic segmentation
AuthorTomislav Pavković
MentorSiniša Šegvić (mentor)
Committee memberSiniša Šegvić (predsjednik povjerenstva)
Committee memberJosip Šarić (član povjerenstva)
Committee memberPetra Bevandić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2022-07-08, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Panoptička segmentacija problem je računalnog vida koji svakom pikselu slike pridružuje oznaku razreda i identifikator primjerka. Ujedinjuje semantičku segmentaciju i segmentaciju primjeraka. U ovom su radu opisani modeli panoptičke segmentacije, njihova podjela i način vrednovanja. Također je dan kratki opis osnova dubokih neuronskih mreže, posebno konvolucijskih mreža, i postupka učenja. Opisani modeli Panoptic SwiftNet, Panoptic DeepLab i Mask2Former naučeni su i vrednovani na skupu podataka Cityscapes. Modeli su uspoređeni na temelju vrijednosti panoptičke kvalitete i vremena zaključivanja. Model Panoptic SwiftNet naučen je na slikama rezolucije 256x512, 512x1024 i 1024x2048. Na temelju tih rezultata, napravljena je analiza utjecaja rezolucije ulaznih slika na vrijednost panoptičke kvalitete i vrijeme zaključivanja modela.
Abstract (english)
Panoptic segmentation is a computer vision task that assigns a class label and instance index to every pixel on an image. It unifies semantic and instance segmentation. This paper describes panoptic segmentation models, their classification, and the method of evaluation. It also provides a short description of deep neural networks basics, especially convolutional networks, and the learning process. Panoptic SwiftNet, Panoptic DeepLab and Mask2Former models are described and evaluated on the Cityscapes dataset. Models are compared based on panoptic quality value and inference time. Panoptic SwiftNet model is learned on images with resolution of 256x512, 512x1024 and 1024x2048. An analysis of the influence that input image resolution has on panoptic quality value and inference time was made based on these results.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:243394
Study programmeTitle: Computing
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 conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2023-01-15 11:00:34