Chinese traditional Visual Cultural Symbols recognition based on SPM muti-feature extraction

Globalization is a historic chance for China development. Moreover, Chinese traditional Visual Cultural Symbols (VCS) is a great tag of China. Image is a carrier of VCS. With the help of machine learning, we can recognize these VCS among numerous images very well. In this paper we propose combining with several features such as SIFT, HOG, RGB, LBP to describe an image and coding these features into higher dimensional vectors. Next, the coded vectors are pooled together. Besides, we also make full use of Bag-of-Words (BOW) and Spatial Pyramid Matching (SPM) theory in order to obtain a more ideal recognition. Experiments show that it's really a feasible approach.