The accuracy (%) of the presented approaches.
收藏数据链接：
官方服务：
资源简介：
Each row corresponds to different descriptor. The first part of the shortcut represents a type of pooling encoder, while the second part is a type of local descriptor. Each column corresponds to different classifier. Original classifier corresponds to linear SVM with C = 1 (the classifier originally used by Cimpoi et al. [1]). The other classifiers use Bayes optimization (see Section Method for details). The best results for each classifier are in bold.
创建时间：
2017-09-15
