Structural analysis of video by histogram-based description of local space-time appearance

doctoral thesis

• Title (croatian): Strukturna analiza videa korištenjem opisa lokalnog prostorno-vremenskog izgleda zasnovanoga na histogramima
• Author: Karla Brkić
• Mentor: Zoran Kalafatić (mentor)
• Mentor: Axel Pinz (komentor); strani drzavljanin: Nije dostupno
• Committee member: Zoran Kalafatić (predsjednik povjerenstva)
• Committee member: Axel Pinz (član povjerenstva); strani drzavljanin: Nije dostupno
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; (Department of Electronics, Microelectronics, Computer and Intelligent Systems); Zagreb
• Granter: Austria: Graz University of Technology; Graz
• Defense date and country: 2013-07-02, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing; Artificial Intelligence
• Universal decimal classification (UDC): 004 - Computer science and technology. Computing. Data processing

Abstract

The methods for representation, classification and reasoning about video data are actively researched in the computer vision community. Different research applications drive different views on what is essentially the same type of information. The goal of this thesis is to investigate the methods for video analysis that are generally applicable to any kind of video data, regardless of the phenomenon represented by the data. Special emphasis is placed on the analysis of videos in an online scenario, where not all frames of a video are available in advance. The thesis introduces the notion of a spatio-temporal structure as a basic unit of video information. A method for representing spatio-temporal structures as grids of histogram is proposed, and the derived grid-of-histograms representation is used to build two different kinds of descriptors: the spatio-temporal appearance (STA) descriptors and the COIN descriptors. The STA descriptors model either the average local appearance or the distributions of local appearance, while the COIN descriptor models change in local appearance by assigning weights to four semantically meaningful hypotheses. A detailed experimental evaluation is provided, illustrating the suitability of STA descriptors as feature vectors in the problems of traffic sign recognition, human action recognition and dynamic texture recognition, and the suitability of COIN descriptors in determining the structural properties of the 3D world.

Abstract (croatian)

Metode za predstavljanje, raspoznavanje i zaključivanje o video podacima aktivno se istražuju u računalnom vidu. U različitim se primjenama najčešće različito gleda na u osnovi istu vrstu informacije. Cilj ove disertacije je istražiti metode za analizu videa koje su općenito primjenjive na bilo koju vrstu video podataka, neovisno o pojavi koju podaci predstavljaju. Poseban je naglasak stavljen na inkrementalnu analizu videa u slučaju kad svi okviri videa nisu dostupni unaprijed. Disertacija uvodi koncept prostorno-vremenske strukture kao osnovne jedinice informacije u videu. Predlaže se metoda za prikaz prostorno-vremenskih struktura mrežama histograma, te se uvedeni prikaz koristi za izgradnju dvije vrste deskriptora: deskriptora prostorno-vremenskog izgleda (STA), te deskriptora COIN. Deskriptori STA modeliraju ili prosječan lokalni izgled ili distribuciju lokalnog izgleda, dok deskriptor COIN modelira promjenu u lokalnom izgledu pridjeljivanjem težina četirima semantičkim hipotezama. Prikazani su rezultati detaljne eksperimentalne evaluacije koja ilustrira primjenjivost deskriptora STA u problemima raspoznavanja prometnih znakova, ljudskih akcija i dinamičkih tekstura, te primjenjivost deskriptora COIN u određivanju strukturnih svojstava 3D svijeta.

• Keywords: video analysis; spatio-temporal appearance; spatio-temporal structure; video descriptors; semantic descriptors; action recognition; dynamic textures
• Keywords (croatian): analiza videa; prostorno-vremenski izgled; prostorno-vremenska struktura; semantički opisnik; raspoznavanje akcija; dinamičke teksture
• Language: english
• URN:NBN: urn:nbn:hr:168:337571
• Project: Number: 036-0361935-1954; Title: Teorija, modeliranje i uporaba autonomno orijentiranih računarskih struktura; Leader: Slobodan Ribarić; Jurisdiction: Croatia; Funder: MZOS; Funding stream: ZP
• Study programme: Title: Computer Science; Study programme type: university; Study level: postgraduate; Academic / professional title: Doktor znanosti (Doktor znanosti)
• Catalog URL: http://lib.fer.hr/cgi-bin/koha/opac-detail.pl?biblionumber=41322
• Type of resource: Text
• Extent: 210 str ; 30 cm
• File origin: Born digital
• Access conditions: Open access
• Created on: 2020-01-16 08:57:52