| Yannis Avrithis |
| Video Sequence Analysis for Content Description, Summarization and Content-Based Retrieval |
| Ph.D. Thesis, School of Electrical and Computer Engineering, National Technical University of Athens |
|
ABSTRACT
|
| The main research area of this Ph.D. thesis is image and video sequence processing and analysis for description and indexing of their visual content. Its objective is to contribute in the development of an automated computational system that has the capabilities of object-based segmentation of audiovisual material, automatic content description and annotation, summarization for preview and browsing, as well as content-based search and retrieval. The thesis consists of four parts. The first part introduces video sequence analysis, segmentation and object extraction based on color, motion, as well as depth field in the case of stereoscopic video sequences. A fusion technique is proposed that combines individual cue segmentations and allows for reliable identification of semantic objects. The second part refers to automatic annotation of the visual content by means of feature vectors calculated by multidimensional fuzzy classification of low-level object descriptors. This information is used for summarization, which is implemented by optimal selection of a limited set of key frames and shots providing meaningful visual content description. The representation of the selected material by feature vectors is then employed for content-based search and retrieval. In the third part, the problem of object contour analysis and representation is examined, with application to shape-based object classification and retrieval. An original contour normalization scheme is presented, permitting invariant shape representation with respect to a large number of transformations without any actual loss of information. In the fourth part, a novel technique is proposed for temporal segmentation and parsing of broadcast news recordings into elementary story units or news topics using visual cues. The technique is based on an advanced algorithm for automatic detection of human faces; the extracted information is also employed for the development of new semantic criteria for content-based retrieval. Finally, conclusions are drawn and a number of issues are proposed that could form the basis for future research.
|
| 13 March , 2001 |
| Yannis Avrithis, "Video Sequence Analysis for Content Description, Summarization and Content-Based Retrieval", Ph.D. Thesis, School of Electrical and Computer Engineering, National Technical University of Athens |
[  PDF] [
BibTex] [
Print] [
 Back] |
