一种基于压缩感知理论的纹理分类方法.

According to the theories of sparse representation and compressed sensing, this paper presented a simple, novel approach for texture classification based on bag-of-words model. At the feature extraction stage, it extracted a small set of random features from local image patches. It embedded the random features into a bag-of-words model to perform texture classification; thus, carried out learning and classification in a compressed domain, yet by leveraging the sparse nature of texture images, our approach outperformed traditional feature extraction methods which involved careful design and complex steps. It conducted extensive experiments on the CUReT databases. Results show that our approach leads to significant improvements in classification accuracy and instantaneity. [ABSTRACT FROM AUTHOR]