Dictionary Learning for Fast Classification Based on Soft-thresholding
Dictionary Learning for Fast Classification Based on Soft-thresholding
复制标题
DOI：
10.1007/s11263-014-0784-7
复制
发表时间：
2014-02
影响因子：
19.5
通讯作者：
Alhussein Fawzi;M. Davies;P. Frossard
中科院分区：
文献类型：
--
作者：
Alhussein Fawzi;M. Davies;P. Frossard
基于稀疏表示的分类器最近被证明在许多视觉识别和分类任务中提供了优异的结果。然而，在测试时计算稀疏表示的高成本是限制这些方法在大规模问题中或在计算能力受限的情况下的适用性的主要障碍。我们认为在本文中，一个简单而有效的替代稀疏编码的特征提取。我们研究了一种分类方案，该方案在字典中应用软阈值非线性映射，然后使用线性分类器。针对这种低复杂度的分类结构，提出了一种新的监督字典学习算法。将字典学习问题转化为一个凸差分（DC）规划，并利用迭代DC求解器有效地解决了字典学习和线性分类器的联合学习问题。我们在几个数据集上进行了实验，并表明我们的学习算法，利用分类问题的结构优于一般的学习过程。我们的简单分类器的基础上软阈值也与最近的稀疏编码分类器，当字典学习适当的竞争。与其他分类器相比，所采用的分类方案在测试阶段还需要更少的计算时间。该方案显示了充分训练的软阈值映射分类的潜力，并为视觉问题的非常有效的分类方法的发展铺平了道路。
Classifiers based on sparse representations have recently been shown to provide excellent results in many visual recognition and classification tasks. However, the high cost of computing sparse representations at test time is a major obstacle that limits the applicability of these methods in large-scale problems, or in scenarios where computational power is restricted. We consider in this paper a simple yet efficient alternative to sparse coding for feature extraction. We study a classification scheme that applies thesoft-thresholdingnonlinear mapping in a dictionary, followed by a linear classifier. A novel supervised dictionary learning algorithm tailored for this low complexity classification architecture is proposed. The dictionary learning problem, whichjointlylearns the dictionary and linear classifier, is cast as adifference of convex(DC) program and solved efficiently with an iterative DC solver. We conduct experiments on several datasets, and show that our learning algorithm that leverages the structure of the classification problem outperforms generic learning procedures. Our simple classifier based on soft-thresholding also competes with the recent sparse coding classifiers, when the dictionary is learned appropriately. The adopted classification scheme further requires less computational time at the testing stage, compared to other classifiers. The proposed scheme shows the potential of the adequately trained soft-thresholding mapping for classification and paves the way towards the development of very efficient classification methods for vision problems.