Image classification is an important problem in computer vision. The sparse coding spatial\npyramid matching (ScSPM) framework is widely used in this field. However, the sparse coding\ncannot effectively handle very large training sets because of its high computational complexity,\nand ignoring the mutual dependence among local features results in highly variable sparse codes\neven for similar features. To overcome the shortcomings of previous sparse coding algorithm,\nwe present an image classification method, which replaces the sparse dictionary with a stable\ndictionary learned via low computational complexity clustering, more specifically, a k-medoids\ncluster method optimized by k-means++. The proposed method can reduce the learning complexity\nand improve the featureâ??s stability. In the experiments, we compared the effectiveness of our method\nwith the existing ScSPM method and its improved versions. We evaluated our approach on two\ndiverse datasets: Caltech-101 and UIUC-Sports. The results show that our method can increase the\naccuracy of spatial pyramid matching, which suggests that our method is capable of improving\nperformance of sparse coding features.
Loading....