prikaz prve stranice dokumenta Analiza videozapisa nogometnih utakmica tehnikama računalnog vida
No public access
master's thesis
Analiza videozapisa nogometnih utakmica tehnikama računalnog vida
Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. urn:nbn:hr:168:333947
Pavić, Tihomir
University of Zagreb
Faculty of Electrical Engineering and Computing

Institutional repository: FER Repository

Cite this document

Pavić, T. (2024). Analiza videozapisa nogometnih utakmica tehnikama računalnog vida (Master's thesis). Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing. Retrieved from https://urn.nsk.hr/urn:nbn:hr:168:333947

Pavić, Tihomir. "Analiza videozapisa nogometnih utakmica tehnikama računalnog vida." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. https://urn.nsk.hr/urn:nbn:hr:168:333947

Pavić, Tihomir. "Analiza videozapisa nogometnih utakmica tehnikama računalnog vida." Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, 2024. https://urn.nsk.hr/urn:nbn:hr:168:333947

Pavić, T. (2024). 'Analiza videozapisa nogometnih utakmica tehnikama računalnog vida', Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, accessed 26 December 2024, https://urn.nsk.hr/urn:nbn:hr:168:333947

Pavić T. Analiza videozapisa nogometnih utakmica tehnikama računalnog vida [Master's thesis]. Zagreb: University of Zagreb, Faculty of Electrical Engineering and Computing; 2024 [cited 2024 December 26] Available at: https://urn.nsk.hr/urn:nbn:hr:168:333947

T. Pavić, "Analiza videozapisa nogometnih utakmica tehnikama računalnog vida", Master's thesis, University of Zagreb, Faculty of Electrical Engineering and Computing, Zagreb, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:168:333947

Metadata
TitleAnaliza videozapisa nogometnih utakmica tehnikama računalnog vida
Title (english)Analysis of soccer match video recordings using computer vision techniques
AuthorTihomir Pavić
MentorTomislav Hrkać (mentor)
Committee memberTomislav Hrkać (predsjednik povjerenstva)
Committee memberSiniša Šegvić (član povjerenstva)
Committee memberZoran Kalafatić (član povjerenstva)
GranterUniversity of Zagreb
Faculty of Electrical Engineering and Computing
Zagreb
Defense date and country2024-07-02, Croatia
Abstract
Cilj ovog rada bio je razviti i ispitati metodu koja može unutar videozapisa cijele nogometne utakmice detektirati ključne nogometne akcije s ciljem kreiranja video sažetaka nogometne utakmice. Metoda se sastoji od dvije konvolucijske neuronske mreže. Prvi dio čini arhitektura ResNet-152 koja iz ulaznog slijeda okvira (engl. frames) na izlazu daje slijed izdvojenih značajki. Sekvence tih značajki ulaze u drugi model imena RMS-Net koji za svaku sekvencu mora odrediti nalazi li se neka od ključnih nogometnih akcija unutar te sekvence radeći klasifikaciju sekvence u jednu od sedam klasa te za njih i predvidjeti relativni vremenski pomak kako bi znali trenutak u kojem se akcija dogodila. Ako model ne pronađe akciju unutar sekvence, tada tu sekvencu klasificira kao klasu pozadine koja ne ulazi u konačni video sažetak.
Abstract (english)
The aim of this thesis was to develop and test a method that can detect key soccer actions within a full match video to create video summaries of the soccer match. The method consists of two convolutional neural networks. The first part is a ResNet-152 architecture, which extracts a sequence of features from the input sequence of frames. The sequences of these features are fed into a second model named RMS-Net, which for each sequence must determine whether any of the key soccer actions are present within that sequence by classifying the sequence into one of seven classes and predicting the relative temporal offset to know the exact moment the action occurred. If the model does not find an action within the sequence, it classifies that sequence as a background class, which is not included in the final video summary.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:168:333947
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 on2024-10-12 21:47:53