文献类型：期刊文献

英文题名：SubMIL: Discriminative subspaces for multi-instance learning

作者：Yuan, Jiazheng[1,2];Huang, Xiankai[3];Liu, Hongzhe[1];Li, Bing[4];Xiong, Weihua[4]

第一作者：Yuan, Jiazheng

通讯作者：Yuan, JZ[1]

机构：[1]Beijing Key Lab Informat Serv Engn, Beijing 100101, Peoples R China;[2]Beijing Union Univ, Comp Technol Inst, Beijing 100101, Peoples R China;[3]Beijing Union Univ, Tourism Inst, Beijing 100101, Peoples R China;[4]Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit NLPR, Beijing 100190, Peoples R China

第一机构：Beijing Key Lab Informat Serv Engn, Beijing 100101, Peoples R China

通讯机构：[1]corresponding author), Beijing Key Lab Informat Serv Engn, Beijing 100101, Peoples R China.

年份：2016

卷号：173

起止页码：1768-1774

外文期刊名：NEUROCOMPUTING

收录：;EI(收录号：20155201737382);Scopus(收录号：2-s2.0-84955202801);WOS:【SCI-EXPANDED(收录号:WOS:000366879800129)】；

基金：This paper is supported by the following projects: the National Natural Science Foundation of China (Nos. 61271369, 61372148, and 61370038), Beijing Natural Science Foundation (4152016, 4152018), the National Key Technology R&D Program (2014BAK08B02, 2015BAH55F03), Funding Project for Academic Human Resources Development in Beijing Union University (BPHR2014A04, BPHR2014E02), and the Project of Construction of Innovative Teams and Teacher Career Development for Universities and Colleges under Beijing Municipality (CIT&TCD 20130513, IDHT 20140508).

语种：英文

外文关键词：Multi-instance learning; Low rank; Subspace

摘要：As an important learning scheme for Multi-Instance Learning (MIL), the Instance Prototype (IP) selection-based MIL algorithms transform bags into a new instance feature space and achieve impressed classification performance. However, the number of IPs in the existing algorithms linearly increases with the scale of the training data. The performance and efficiencies of these algorithms are easily limited by the high dimension and noise when facing a large scale of training data. This paper proposes a discriminative subspaces-based instance prototype selection method that is suitable for reducing the computation complexity for large scale training data. In the proposed algorithm, we introduce the low-rank matrix recovery technique to find two discriminative and clean subspaces with less noise; then present a l(2,1) norm-based self-expressive sparse coding model to select the most representative instances in each subspace. Experimental results on several data sets show that our algorithm achieves superior and stable performance but with lower dimension compared with other IP selection strategies. (C) 2015 Elsevier B.V. All rights reserved.