自适应多视角学习及其在图像分类中应用.docVIP

自适应多视角学习及其在图像分类中应用 　　文章编号:10019081(2013)07195505 　　doi:10.11772/j.issn.10019081.2013.07.1955 　　摘 要： 　　针对现有多视角学习算法在构建近邻图时缺乏数据自适应性问题，提出一种自适应多视角学习(AMVL)算法。该算法首先利用L1范数具有自动数据样本选择的特性，对不同视角分别构建有向的L1图；然后根据得到的L1图，最小化各个视角下的低维重建误差；最后对不同视角间进行多视角全局坐标对齐，得到自适应多视角学习算法的目标函数。此外，还提出一种迭代优化求解方法来对所提目标函数进行优化求解。将该算法应用到图像分类问题，在Corel5K和NUSWIDEOBJECT两个公共图像数据集上与现有算法进行对比。实验结果表明：所提方法在这两个数据集上可以分别提高最高5%和2%的分类准确率；优化求解算法可以保证在100次迭代内收敛；算法所得到的近邻数目具有数据自适应性。 　　关键词：多视角学习；自适应；L1图；图像分类 　　中图分类号：TP391.413 　　文献标志码:A 　　英文标题 　　Adaptive multiview learning and its application to image classification 　　英文作者名 　　MAO Jinlian* 　　英文地址( 　　Institute of Information Technology, Zhejiang Business College, Hangzhou Zhejiang 310053, China英文摘要) 　　Abstract： 　　Since the existing multiview learning algorithms often suffer from the problem of lacking data adaptiveness in the nearest neighbor graph construction procedure, an Adaptive MultiView Learning (AMVL) algorithm was proposed. Firstly, by utilizing the automatic data sample selection property of L1 norm constraint, multiple viewrelated directed L1graphs were constructed. Secondly, according to the obtained L1 graphs, the algorithm tried to minimize the low dimensional reconstruction error in each view. Lastly, the objective function of the proposed adaptive multiview learning algorithm was obtained by performing global coordinate alignment process in different views. Moreover, an iterative optimization method was also proposed to solve the proposed objective function. The algorithm was applied to the problem of image classification on two public image datasets, i.e., Corel5K and NUSWIDEOBJECT, and compared with several existing methods. The experimental results show that： a) the proposed algorithm can increase the classification accuracy up to 5% and 2% respectively on these two datasets; b) the optimization method convergences within 100 iterations; c) the number of nearest neighborhoods learned by the algorithm is datumadaptive