Low-rank matrix regression for image feature extraction and feature selection

In many image processing and pattern recognition problems, the input data is commonly the images. The image could be represented as the matrix form. The natural structure information of the matrix is useful for data analysis and representation. However, most of existing methods commonly convert the image as a vector form, which destroys the natural structure of the image. To fully utilize this kind of structure information, we propose a low-rank matrix regression model for feature extraction and feature selection. To efficiently solve the objective functions of the proposed methods, we develop an optimization algorithm based on the alternating direction method of multipliers method. Promising experimental results have demonstrated the effectiveness of our proposed methods. (C) 2020 Elsevier Inc. All rights reserved.