Mutual Information-based Multi-channel Joint Sparse Model for Histopathological Images Classification

In the traditional joint sparse model, a dictionary was used for feature representation of the common or unique components. This leads to the low discrimination of the sparse coding coefficients. In this paper, mutu-al information-based multi-channel joint sparse model is proposed for histopathological image classification. The training samples are clustered into R, G and B channel dictionaries by using K-means. By exploring the mutual information between training samples and three dictionaries, the irrelevant atoms are deleted, meanwhile, a shared dictionary and three unique dictionaries constructed. Simultaneously, multi-channel joint sparse model is designed based on the shared dictionary and three unique dictionaries. Furthermore, in order to represent image feature of different levels, the spatial pyramid matching is used to the multi-channel joint sparse coding. Finally, the joint sparse coding coefficients are used to train the SVM for histopathological images classification. The experimental results show that the proposed model has power feature representation ability and improve greatly the discrimination of coding coefficients. Thus the better classification performance and the power robustness can be obtained with compared to the traditional models. © 2018, Beijing China Science Journal Publishing Co. Ltd. All right reserved.