Combining Local Feature Histograms of Different Granularities

Authors:
Ville Viitaniemi;Jorma Laaksonen
Affiliations:
Department of Information and Computer Science, Helsinki University of Technology, TKK, Finland FI-02015;Department of Information and Computer Science, Helsinki University of Technology, TKK, Finland FI-02015
Venue:
SCIA '09 Proceedings of the 16th Scandinavian Conference on Image Analysis
Year:
2009

Citing 5
Cited 0

Scale & Affine Invariant Interest Point Detectors

International Journal of Computer Vision
Distinctive Image Features from Scale-Invariant Keypoints

International Journal of Computer Vision
The Pyramid Match Kernel: Efficient Learning with Sets of Features

The Journal of Machine Learning Research
Experiments on Selection of Codebooks for Local Image Feature Histograms

VISUAL '08 Proceedings of the 10th international conference on Visual Information Systems: Web-Based Visual Information Search and Management
Improving the accuracy of global feature fusion based image categorisation

SAMT'07 Proceedings of the semantic and digital media technologies 2nd international conference on Semantic Multimedia

Quantified Score

Hi-index	0.00

Visualization

Abstract

Histograms of local features have proven to be powerful representations in image category detection. Histograms with different numbers of bins encode the visual information with different granularities. In this paper we experimentally compare techniques for combining different granularities in a way that the resulting descriptors can be used as feature vectors in conventional vector space learning algorithms. In particular, we consider two main approaches: fusing the granularities on SVM kernel level and moving away from binary or hard to soft histograms. We find soft histograms to be a more effective approach, resulting in substantial performance improvement over single-granularity histograms.