Post-graduate theses
Search command : Author="Φλουρής"
And Author="Γεώργιος"
Current Record: 690 of 824
|
|
| Identifier |
uch.csd.msc//1999grinias |
| Title |
Ανάλυση Περιεχομένου Video |
| Alternative Title |
Video Content Analysis |
| Creator |
Grinias, Elias K
|
| Abstract |
Nowadays, the progress in the development of broadband networks which is followed by the concurrent progress in building high-powered workstations, has increased the data manipulation capabilities. This advent, provides the ability to efficiently parse the multimedia information which in turn, increases the importance of efficient optical information management systems. The management consists of video and image manipulation methods, which facilitate the decreasing of the data storage capacity as well as the efficient indexing and retrieval of their content. The special nature of the optical information compared to other digital data sources, such as the text, has lead to the development of indexing/retrieval methods that are based on the optical information content. In this thesis, a feature extraction system is described which is based on the content of the video and its images. The design of the system follows the current approaches on the indexing/retrieval and processing of the optical information. In the first part of the thesis, a fast video segmentation procedure to scenes, is presented. The method cooperates with a key-frame extraction procedure in order to fill the entries of a hierarchical data model, which in general, follows the video indexing data model of the stand ard MPEG7. The method has been applied on MPEG compressed video. In the second part of the thesis, two object extraction methods are presented. The first one is based on the object color, while the second uses the object motion information, to segment the image. Both of them, are based on the segmentation algorithm Seeded Region Growing (SRG). First, the extension of the user-guided SRG algorithm is described, in order to be used in the color image segmentation. Then, an automatic region extraction method which is based on the color of images is proposed. The algorithm is performed on images with faces. The results include the comparison between the systems GLHS and Lab. The extracted regions may then be used as input to the motion segmentation and tracking system.
|
| Issue date |
1999-11-01 |
| Date available |
1999-10-20 |
| Collection
|
School/Department--School of Sciences and Engineering--Department of Computer Science--Post-graduate theses
|
|
|
Type of Work--Post-graduate theses
|
| Permanent Link |
https://elocus.lib.uoc.gr//dlib/6/8/a/metadata-dlib-1999grinias.tkl
|
| Views |
494 |
Collections
