摘要 : These days, various forms of data can be seen in real word. Images as a popular type of data are widely used in many resources such as web pages, video, etc. Meanwhile, each images as a source of data is comprised of many features... 展开 These days, various forms of data can be seen in real word. Images as a popular type of data are widely used in many resources such as web pages, video, etc. Meanwhile, each images as a source of data is comprised of many features. Different methods of feature extraction present occasionally different descriptions of the data, each can be considered as one view of data. In this study, K-Singular Value Decomposition (K-SVD) method which is a popular method for sparse dictionary learning is developed for multi view data. Sparse dictionary learning has been applied in multi view data identification problem to find out which classes belongs to basis vector. Different feature extraction methods such as Scale-Invariant Feature Transform (SIFT) and GIST of a scene are used for plant identification to describe images of leaves in many studies, but in most of them, these different views of data are merged as high dimensional features. In this study K-SVD, as a dimension reduction algorithm is developed for sparse dictionary learning for multi view data create low dimensional visualization of images in plant identification scope. The experimental results show improvement in accuracy compared to regular K-SVD and early fusion of features. 收起