Laplacian sparse dictionary learning for image classification based on sparse representation  

作　　者：Fang LI Jia SHENG San-yuan ZHANG 

机构地区：[1]College of Computer Science and Technology, Zhejiang University, Hangzhou 310027, China [2]School of Computer Science and Information Security, Guilin University of Electronic Technology, Cruilin 541004, China [3]Guangxi Key Laboratory of Trusted Software, Guilin University of Electronic Technology, Cruilin 541004, China

出　　处：《Frontiers of Information Technology & Electronic Engineering》2017年第11期1795-1805,共11页信息与电子工程前沿（英文版）

基　　金：Project supported by the National Natural Science Foundation of China （Nos. 61272304 and 61363029） and the Guangxi Key Laboratory of Trusted Software （No. kx201313）

摘　　要：Sparse representation is a mathematical model for data representation that has proved to be a powerful tool for solving problems in various fields such as pattern recognition, machine learning, and computer vision. As one of the building blocks of the sparse representation method, dictionary learning plays an important role in the minimization of the reconstruction error between the original signal and its sparse representation in the space of the learned dictionary. Although using training samples directly as dictionary bases can achieve good performance, the main drawback of this method is that it may result in a very large and inef- ficient dictionary due to noisy training instances. To obtain a smaller and more representative dictionary, in this paper, we propose an approach called Laplacian sparse dictionary （LSD） learning. Our method is based on manifold learning and double sparsity. We incorporate the Laplacian weighted graph in the sparse representation model and impose the 11-norm sparsity on the dictionary. An LSD is a sparse overcomplete dictionary that can preserve the intrinsic structure of the data and learn a smaller dictionary for each class. The learned LSD can be easily integrated into a classification framework based on sparse representation. We compare the proposed method with other methods using three benchmark-controlled face image databases, Extended Yale B, ORL, and AR, and one uncontrolled person image dataset, i-LIDS-MA. Results show the advantages of the proposed LSD algorithm over state-of-the-art sparse representation based classification methods.稀疏表示作为数据表示的一种数学模型,是解决模式识别、机器学习、计算机视觉等领域问题的有力工具。字典学习是稀疏表示方法的重要组成部分,在对原始信号及其在字典学习空间中的重建误差的最小化上发挥着重要的作用。在稀疏表示模型中,直接利用训练样本作为字典可以取得良好的性能。但由于训练样本含有噪声,这样的字典很大且效率低下。为取得更小且表现更好的字典,本文提出一种基于流形学习及双稀疏理论的拉普拉斯稀疏字典学习方法(Laplacian sparse dictionary,LSD)。本文将拉普拉斯权重图加入稀疏表示的模型,并对字典加以l_1范数约束。LSD是一个稀疏的过完备字典,可保持数据的内在结构,并为每个类学习一个更小的字典。学习得到的字典可以嵌入基于稀疏表示的分类框架。将本文提出的方法和其它方法在三个基准的约束人脸数据(Extended Yale B、ORL、AR)和一个无约束的行人数据图像数据库i-LIDS-MA上进行对比实验。结果显示本文提出的LSD算法比当前基于分类的稀疏表示的方法更有优势。

关 键 词：Sparse representation Laplacian regularizer Dictionary learning Double sparsity MANIFOLD 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]