Blobworld: A system for region-based image indexing and retrieval
Research output: Contribution to journal › Conference article › Research › peer-review
Blobworld is a system for image retrieval based on finding coherent image regions which roughly correspond to objects. Each image is automatically segmented into regions (“blobs”) with associated color and texture descriptors. Queryingi s based on the attributes of one or two regions of interest, rather than a description of the entire image. In order to make large-scale retrieval feasible, we index the blob descriptions using a tree. Because indexing in the high-dimensional feature space is computationally prohibitive, we use a lower-rank approximation to the high-dimensional distance. Experiments show encouraging results for both querying and indexing.
Original language | English |
---|---|
Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
Pages (from-to) | 509-517 |
Number of pages | 9 |
ISSN | 0302-9743 |
DOIs | |
Publication status | Published - 1999 |
Externally published | Yes |
Event | 3rd International Conference on Visual Information Systems, VISUAL 1999 - Amsterdam, Netherlands Duration: 2 Jun 1999 → 4 Jun 1999 |
Conference
Conference | 3rd International Conference on Visual Information Systems, VISUAL 1999 |
---|---|
Country | Netherlands |
City | Amsterdam |
Period | 02/06/1999 → 04/06/1999 |
Sponsor | Advanced School for Computing and Imaging, Netherlands Computers Science Research Foundation, Royal Academy of Arts and Sciences, Shell Nederland B.V, University of Amsterdam |
Bibliographical note
Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 1999.
ID: 302060207