Generalized Histogram Intersection kernel for image classification

Kernel-based Support Vector Machine (SVM) is widely used in many fields (e.g. image classification) for its good generalization, in which the key factor is to design effective kernel functions based on efficient features. In this paper, we propose a new approach that uses a combination of global and local image features to represent images and learns Support Vector Machine classifier with a new and fast kernel, which is named Generalized Histogram Intersection (GHI) kernel. We then conduct a comparative evaluation with several state-of- the-art recognition methods on two popular benchmark datasets (Corel1K and Caltech101). The results show our algorithm to be more accurate than current approaches.