Finding pictures of objects in large collections of images
Research output: Contribution to journal › Conference article › Research › peer-review
Retrieving images from very large collections, using image content as a key, is becoming an important problem. Users prefer to ask for pictures using notions of content that are strongly oriented to the presence of abstractly defined objects. Computer programs that implement these queries automatically are desirable, but are hard to build because conventional object recognition techniques from computer vision cannot recognize very general objects in very general contexts. This paper describes our approach to object recognition, which is structured around a sequence of increasingly specialized grouping activities that assemble coherent regions of image that can be shown to satisfy increasingly stringent constraints. The constraints that are satisfied provide a form of object classification in quite general contexts. This view of recognition is distinguished by: far richer involvement of early visual primitives, including color and texture; hierarchical grouping and learning strategies in the classification process; the ability to deal with rather general objects in uncontrolled configurations and contexts. We illustrate these properties with four case-studies: one demonstrating the use of color and texture descriptors; one showing how trees can be described by fusing texture and geometric properties; one learning scenery concepts using grouped features; and one showing how this view of recognition yields a program that can tell, quite accurately, whether a picture contains naked people or not.
Original language | English |
---|---|
Journal | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
Pages (from-to) | 335-360 |
Number of pages | 26 |
ISSN | 0302-9743 |
DOIs | |
Publication status | Published - 1996 |
Externally published | Yes |
Event | International Workshop on Object Representation in Computer Vision II, ECCV 1996 - Cambridge, United Kingdom Duration: 13 Apr 1996 → 14 Apr 1996 |
Conference
Conference | International Workshop on Object Representation in Computer Vision II, ECCV 1996 |
---|---|
Country | United Kingdom |
City | Cambridge |
Period | 13/04/1996 → 14/04/1996 |
Bibliographical note
Funding Information:
We would like to thank R. Blasi and K. Murphy who collaborated with S. Belongie in the work on learning decision trees for visual concept classification. We thank Joe Mundy for suggesting that the response of a grouper may indicate the presence of an object. Aspects of this research were supported by the National Science Foundation under grants IRI-9209728, IRI-9420716, IRI-9501493, an NSF Young Investigator award, an NSF Digital Library award IRI-9411334, an instrumentation award CDA-9121985, and by a Berkeley Fellowship.
Publisher Copyright:
© 1996, Springer Verlag. All rights reserved.
ID: 302163653