期刊名称：International Journal of Electronics, Communication and Soft Computing Science and Engineering
印刷版ISSN：2277-9477
出版年度：2015
卷号：4
期号：Special 3
出版社：IJECSCSE
摘要：Subspace learning is an effective image feature extraction and classification technique, which generally includes supervised and unsu pervised subspace learning methods. Parallel computing is an effective technique in solving large - scale learning problems. It splits the original task into several subtasks and reduces the time complexity. Parallel computing reduces the time complexity by splitting the original task into several subtasks. a parallel subspace learning framework is developed in this paper. In this method, first the sample set is divided into several subsets by designing two random data division strategies that are equal data division and unequal data division. These two strategies correspond to equal and unequal computational abilities of nodes under parallel computing environment. Then, projection vectors from each subset are calculated in parallel. The graph embedding techni que is employed to provide a general formulation for parallel feature extraction. After combining the extracted features from all nodes, a unified criterion is presented to select most distinctive features for classification. The proposed approaches can ou tperform several related supervised and unsupervised subspace learning methods, and significantly reduce the computational time.
