2015/02/10 by Soheil Bahrampour, Nasser M. Nasrabadi, Bahrampour, Soheil +5
Engineering · Physics and Astronomy · #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Optical and Acousto-Optic Technologies #Remote-Sensing Image Classification #Sparse and Compressive Sensing Techniques
paper · pdf · doi:10.48550/arxiv.1502.03126
openalex publication_date 2015/02/10 · openalex created_date 2022/10/02 · openalex updated_date 2026/07/28
Dictionary learning algorithms have been successfully used in both\nreconstructive and discriminative tasks, where the input signal is represented\nby a linear combination of a few dictionary atoms. While these methods are\nusually developed under \ℓ1 sparsity constrain (prior) in the input\ndomain, recent studies have demonstrated the advantages of sparse\nrepresentation using structured sparsity priors in the kernel domain. In this\npaper, we propose a supervised dictionary learning algorithm in the kernel\ndomain for hyperspectral image classification. In the proposed formulation, the\ndictionary and classifier are obtained jointly for optimal classification\nperformance. The supervised formulation is task-driven and provides learned\nfeatures from the hyperspectral data that are well suited for the\nclassification task. Moreover, the proposed algorithm uses a joint\n(\ℓ12) sparsity prior to enforce collaboration among the neighboring\npixels. The simulation results illustrate the efficiency of the proposed\ndictionary learning algorithm.\n