Comparing images using color coherence vectors
MULTIMEDIA '96 Proceedings of the fourth ACM international conference on Multimedia
Distinctive Image Features from Scale-Invariant Keypoints
International Journal of Computer Vision
International Journal of Computer Vision
Scene Classification Using a Hybrid Generative/Discriminative Approach
IEEE Transactions on Pattern Analysis and Machine Intelligence
Overview of the CLEF 2009 large-scale visual concept detection and annotation task
CLEF'09 Proceedings of the 10th international conference on Cross-language evaluation forum: multimedia experiments
Hi-index | 0.00 |
This paper describes a method that learns a variety of features to perform photo annotation. We introduce concept-specific regional features and combine them with global features. The regional features were extracted through a novel region selection algorithm based on Multiple Instance Learning. Supervised classification for photo annotation was learned using Support Vector Machines with extended Gaussian Kernels over the χ2 distance, together with a simple greedy feature selection. The method was evaluated using the ImageCLEF 2009 Photo Annotation task and competitive benchmarking results were achieved.