Persistent Link:
http://hdl.handle.net/10150/195757
Title:
Matching Slides to Presentation Videos
Author:
Fan, Quanfu
Issue Date:
2008
Publisher:
The University of Arizona.
Rights:
Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.
Abstract:
Video streaming is becoming a major channel for distance learning (or e-learning). A tremendous number of videos for educational purpose are capturedand archived in various e-learning systems today throughout schools, corporations and over the Internet. However, making information searchable and browsable, and presenting results optimally for a wide range of users and systems, remains a challenge.In this work two core algorithms have been developedto support effective browsing and searching of educational videos. The first is a fully automatic approach that recognizes slides in the videowith high accuracy. Built upon SIFT (scale invariant feature transformation) keypoint matching using RANSAC (random sample consensus), the approach is independent of capture systems and can handle a variety of videos with different styles and plentiful ambiguities. In particular, we propose a multi-phase matching pipeline that incrementally identifies slides from the easy ones to the difficult ones. We achieve further robustness by using the matching confidence as part of a dynamic Hidden Markov model (HMM) that integrates temporal information, taking camera operations into account as well.The second algorithm locates slides in the video. We develop a non-linear optimization method (bundle adjustment) to accurately estimate the projective transformations (homographies) between slides and video frames. Different from estimating homography from a single image, our method solves a set of homographies jointly in a frame sequence that is related to a single slide.These two algorithms open up a series of possibilities for making the video content more searchable, browsable and understandable, thus greatly enriching the user's learning experience. Their usefulness has been demonstrated in the SLIC (Semantically Linking Instructional Content) system, which aims to turnsimple video content into fully interactive learning experience for students and scholars.
Type:
text; Electronic Dissertation
Keywords:
computer vision; video browsing; indexing; slides
Degree Name:
PhD
Degree Level:
doctoral
Degree Program:
Computer Science; Graduate College
Degree Grantor:
University of Arizona
Advisor:
Barnard, Kobus
Committee Chair:
Barnard, Kobus

Full metadata record

DC FieldValue Language
dc.language.isoENen_US
dc.titleMatching Slides to Presentation Videosen_US
dc.creatorFan, Quanfuen_US
dc.contributor.authorFan, Quanfuen_US
dc.date.issued2008en_US
dc.publisherThe University of Arizona.en_US
dc.rightsCopyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.en_US
dc.description.abstractVideo streaming is becoming a major channel for distance learning (or e-learning). A tremendous number of videos for educational purpose are capturedand archived in various e-learning systems today throughout schools, corporations and over the Internet. However, making information searchable and browsable, and presenting results optimally for a wide range of users and systems, remains a challenge.In this work two core algorithms have been developedto support effective browsing and searching of educational videos. The first is a fully automatic approach that recognizes slides in the videowith high accuracy. Built upon SIFT (scale invariant feature transformation) keypoint matching using RANSAC (random sample consensus), the approach is independent of capture systems and can handle a variety of videos with different styles and plentiful ambiguities. In particular, we propose a multi-phase matching pipeline that incrementally identifies slides from the easy ones to the difficult ones. We achieve further robustness by using the matching confidence as part of a dynamic Hidden Markov model (HMM) that integrates temporal information, taking camera operations into account as well.The second algorithm locates slides in the video. We develop a non-linear optimization method (bundle adjustment) to accurately estimate the projective transformations (homographies) between slides and video frames. Different from estimating homography from a single image, our method solves a set of homographies jointly in a frame sequence that is related to a single slide.These two algorithms open up a series of possibilities for making the video content more searchable, browsable and understandable, thus greatly enriching the user's learning experience. Their usefulness has been demonstrated in the SLIC (Semantically Linking Instructional Content) system, which aims to turnsimple video content into fully interactive learning experience for students and scholars.en_US
dc.typetexten_US
dc.typeElectronic Dissertationen_US
dc.subjectcomputer visionen_US
dc.subjectvideo browsingen_US
dc.subjectindexingen_US
dc.subjectslidesen_US
thesis.degree.namePhDen_US
thesis.degree.leveldoctoralen_US
thesis.degree.disciplineComputer Scienceen_US
thesis.degree.disciplineGraduate Collegeen_US
thesis.degree.grantorUniversity of Arizonaen_US
dc.contributor.advisorBarnard, Kobusen_US
dc.contributor.chairBarnard, Kobusen_US
dc.contributor.committeememberAmir, Arnonen_US
dc.contributor.committeememberEfrat, Alonen_US
dc.contributor.committeememberMoon, Bongkien_US
dc.identifier.proquest2597en_US
dc.identifier.oclc659749595en_US
All Items in UA Campus Repository are protected by copyright, with all rights reserved, unless otherwise indicated.