Abstract
Relevance feedback schemes using linear/quadratic estimators have been applied in content-based image retrieval to significantly improve retrieval performance. One major difficulty in relevance feedback is to estimate the support of target images in high dimensional feature space with a relatively small number of training samples. In this paper, we develop a novel scheme based on one-class SVM, which fits a tight hyper-sphere in the nonlinearly transformed feature space to include most of the target images based on the positive examples. The use of kernel provides us an elegant way to deal with nonlinearity in the distribution of the target images, while the regularization term in SVM provides good generalization ability. To validate the efficacy of the proposed approach, we test it on both synthesized data and real-world images. Promising results are achieved in both cases.
Original language | English (US) |
---|---|
Pages | 34-37 |
Number of pages | 4 |
State | Published - 2001 |
Externally published | Yes |
Event | IEEE International Conference on Image Processing (ICIP) 2001 - Thessaloniki, Greece Duration: Oct 7 2001 → Oct 10 2001 |
Other
Other | IEEE International Conference on Image Processing (ICIP) 2001 |
---|---|
Country/Territory | Greece |
City | Thessaloniki |
Period | 10/7/01 → 10/10/01 |
ASJC Scopus subject areas
- Hardware and Architecture
- Computer Vision and Pattern Recognition
- Electrical and Electronic Engineering