prikaz prve stranice dokumenta Reidentifikacija objekata kod praćenja više pokretnih objekata
No public access
master's thesis
Reidentifikacija objekata kod praćenja više pokretnih objekata
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2022. urn:nbn:hr:168:103131
Kalenski, Tea
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Kalenski, T. (2022). Reidentifikacija objekata kod praćenja više pokretnih objekata (Master's thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:103131

Kalenski, Tea. "Reidentifikacija objekata kod praćenja više pokretnih objekata." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2022. https://urn.nsk.hr/urn:nbn:hr:168:103131

Kalenski, Tea. "Reidentifikacija objekata kod praćenja više pokretnih objekata." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2022. https://urn.nsk.hr/urn:nbn:hr:168:103131

Kalenski, T. (2022). 'Reidentifikacija objekata kod praćenja više pokretnih objekata', Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 03 June 2024, https://urn.nsk.hr/urn:nbn:hr:168:103131

Kalenski T. Reidentifikacija objekata kod praćenja više pokretnih objekata [Master's thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2022 [cited 2024 June 03] Available at: https://urn.nsk.hr/urn:nbn:hr:168:103131

T. Kalenski, "Reidentifikacija objekata kod praćenja više pokretnih objekata", Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2022. Available at: https://urn.nsk.hr/urn:nbn:hr:168:103131

Metadata
TitleReidentifikacija objekata kod praćenja više pokretnih objekata
Title (english)Object Re-identification in Multiple Object Tracking
AuthorTea Kalenski
MentorTomislav Pribanić (mentor)
Committee memberTomislav Pribanić (predsjednik povjerenstva)
Committee memberMarko Subašić (član povjerenstva)
Committee memberŽeljka Lučev Vasić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2022-07-15, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
U ovom radu ćemo pokušati zamijeniti konvolucijsku mrežu sa transformerom sa ciljem ostvarivanja bržih i preciznijih rezultata na MOT zadatku. MOT zadatak ima niz izazova poput detekcije objekta, predviđanja pokreta objekta, izrada modela koji daje diskriminativne mape značajki za objekte u svrhu očuvanja identiteta kroz sekvence. Usprkos uspješnom treniranju snažnijeg modela za reidentifikaciju objekta, rezultati metrika praćenja više pokretnih objekata ostaju isti, što nas dovodi do dva moguća zaključka; ili Transformeri još nisu dorasli ovakvom zdatku ili postoji određena granica koliko se može postići samim modelom koji na temelju izgleda reidentificira objekte.
Abstract (english)
In this thesis, we will try to replace a convolution neural network with a transformer in order to achieve faster and more precise results on an MOT task. MOT task has a row of challenges like object detection, object movement prediction, developing a model that provides discriminant feature maps for objects for the purpose of preserving the identity through sequences. Despite successfully training a stronger model for reidentification of an object, metric results for following multiple moving objects stayed the same as before the replacement, which leads us to two possible conclusions; one is that the Transformers are not yet up to such a task, and the second option is that there is a certain limit to what can be achieved by a model that reidentifies objects based on appearance.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:103131
Study programmeTitle: Computing
Study programme type: university
Study level: graduate
Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
Type of resourceText
File originBorn digital
Access conditionsClosed access
Terms of useLicence In copyright icon
Public note
Created on2023-01-13 15:14:16