The nature of statistical learning theory
The nature of statistical learning theory
Machine Learning
Texture Features for Browsing and Retrieval of Image Data
IEEE Transactions on Pattern Analysis and Machine Intelligence
Support Vector Machines for 3D Object Recognition
IEEE Transactions on Pattern Analysis and Machine Intelligence
NETRA: a toolbox for navigating large image databases
NETRA: a toolbox for navigating large image databases
IEEE Transactions on Pattern Analysis and Machine Intelligence
Network Performance Assessment for Neurofuzzy Data Modelling
IDA '97 Proceedings of the Second International Symposium on Advances in Intelligent Data Analysis, Reasoning about Data
Training Support Vector Machines: an Application to Face Detection
CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Texture Features and Learning Similarity
CVPR '96 Proceedings of the 1996 Conference on Computer Vision and Pattern Recognition (CVPR '96)
Empirical Evaluation of Dissimilarity Measures for Color and Texture
ICCV '99 Proceedings of the International Conference on Computer Vision-Volume 2 - Volume 2
Content-based texture image retrieval using fuzzy class membership
Pattern Recognition Letters
| Hi-index | 0.00 |
A novel algorithm is proposed to learn pattern similarities for texture image retrieval. Similar patterns in different texture classes are grouped into a cluster in the feature space. Each cluster is isolated from others by an enclosed boundary, which is represented by several support vectors and their weights obtained from a statistical learning algorithm called support vector machine (SVM). The signed distance of a pattern to the boundary is used to measure its similarity. Furthermore, the patterns of different classes within each cluster are separated by several sub-boundaries, which are also learned by the SVMs. The signed distances of the similar patterns to a particular sub-boundary associated with the query image are used for ranking these patterns. Experimental results on the Brodatz texture database indicate that the new method performs significantly better than the traditional Euclidean distance based approach.