Parcijalni deformabilni modeli za detekciju sudionika u prometu iz video snimaka

master's thesis

• Title (english): Detection of traffic participants in video sequences using deformable part models
• Author: Tomislav Kukolja
• Mentor: Marko Subašić (mentor)
• Committee member: Marko Subašić (predsjednik povjerenstva)
• Committee member: Sven Lončarić (član povjerenstva)
• Committee member: Zvonko Kostanjčar (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2015-07-13, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Rad se bavi analizom te opisom algoritma detektora temeljenog na parcijalnim deformabilnim modelima implementiranog pomoću gramatičkih modela. Cilj rada je postići detekciju više različitih objekata na istoj ulaznoj slici te naći način na koji se algoritam može ubrzati da bi se mogao koristiti za detekciju objekata u video snimkama. Korišten je grafički procesor kao sredstvo za ubrzanje algoritma pri čemu se implementacija ostvarila preko CUDA platforme za paralelno računanje. U radu se iznose problemi koji su se pojavili tijekom izrade rješenja te se daje usporedna analiza vremena izvršavanja u odnosu na originalnu implementaciju. Kao objekti detekcije se u svim testnim primjerima koriste sudionici u prometu: bicikli, pješaci i automobili.

Abstract (english)

This paper deals with the analysis and description of the detector algorithm based on deformable part models implemented using grammar models. The aim of the paper is to achieve the detection of different objects on same input image and find a way to speed up the algorithm that could be used to detect objects in video sequences. Graphics processor is used to speed up the algorithm whereby the implementation is achieved using CUDA toolkit for parallel computing. This paper presents the problems which have appeared during developement of solutions and gives analysis of execution time compared to the original implementation. As objects of detection, in all test examples, are used traffic participants such as bicycles, pedestrians and cars.

• Keywords: detekcija objekata; parcijalni deformabilni modeli; gramatički modeli; grafički procesor; video snimke
• Keywords (english): object detection; deformable part models; grammar models; graphics processing unit; video sequences
• Language: croatian
• URN:NBN: urn:nbn:hr:168:562781
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 09:59:00