|
|
论文:2017,Vol:35,Issue(4):690-697 |
|
|
引用本文： |
|
|
杨红红, 曲仕茹, 金红霞. 基于包级空间多示例稀疏表示的图像分类算法[J]. 西北工业大学学报 |
|
|
Yang Honghong, Qu Shiru, Jin Hongxia. Image Classification Algorithm Based on Bag-Level Space Multiple Instance Learning with Sparse Representation[J]. Northwestern polytechnical university |
|
|
|
|
|
|
|
| 基于包级空间多示例稀疏表示的图像分类算法 |
|
| 杨红红, 曲仕茹, 金红霞 |
|
| 西北工业大学 自动化学院, 陕西 西安 710072 |
| 摘要: |
| 基于多示例学习框架的图像分类算法以其特有的多义性对象表示能力在图像分类中表现出较好的分类效果。但传统的包级空间多示例学习算法在特征选择过程中存在忽略小目标概念区域且包含大量冗余信息的问题，造成部分训练包信息损失，影响分类性能。为此，基于多示例学习与稀疏编码理论提出1种改进的多示例图像分类算法。该算法首先根据同类样本示例聚为一簇的特性，应用聚类算法构造每类图像的视觉词汇，并利用负包中所有示例都为负的特性，对视觉词汇进行约束，消除冗余信息；依据训练样本示例与视觉词汇的相似度，获得每类训练样本的包特征向量。然后，基于稀疏编码理论，对训练包中的包特征向量进行稀疏编码，获得每1类训练样本的字典矩阵。最后，对待分类样本特征进行稀疏线性组合，预测待分类样本的类别标签。通过对COREL数据集图像进行测试，结果表明，与其他多示例学习算法相比，文中提出的方法能较好地解决图像分类问题，具有较高的分类精度。 |
| 关键词:
包特征向量
稀疏表示
多示例学习
图像分类
|
|
| Image Classification Algorithm Based on Bag-Level Space Multiple Instance Learning with Sparse Representation |
|
| Yang Honghong, Qu Shiru, Jin Hongxia |
|
| College of Automation, Northwestern Polytechnical University, Xi'an 710072, China |
| Abstract: |
| The classification algorithm based on multiple instance learning(MIL) has a good performance due to the MIL has disambiguate ability. However, the bag-level space multiple instance learning algorithms always ignore the small target region and contains a large amount of redundant information during feature selection, which may cause the information loss for partial bags and can affect the performance of classification. In this paper, we proposed an improved multiple instance learning classification algorithm based on the framework of multiple instance learning and the sparse coding. Firstly, according to the characteristics of similar samples can cluster into one class, k-means algorithm is used to construct the visual vocabulary for each class of images. To eliminate redundant information, the negative characteristic of negative samples in negative bags is used to constrain the visual vocabulary. The bag feature vectors for each class of training samples are achieved by computing the similarity between the training sample and the visual vocabulary. Then, sparse coding is used to achieve the dictionary matrix for each class of the training samples. Finally, the labels for test images are predicted by linear combination of the dictionary and coefficient, which is achieved in training data, to represent the bag-level features for test images. Experimental results on COREL image data show that the proposed algorithm can better solve the problems in multiple instance learning image classification and achieve higher classification accuracy compared with the other multiple instance learning based image classification algorithms. |
| Key words:
bag feature
sparse representation
multiple instance learning
image classification
|
|
| 收稿日期: 2016-09-29
修回日期:
|
| DOI: |
| 基金项目: 航天科技创新基金（CASC201104）与航空科学基金（2012ZC53043）资助 |
| 通讯作者:
Email： |
| 作者简介: 杨红红(1988—),女,西北工业大学博士研究生,主要从事图像处理与模型识别研究。
|
|
|
|
|
|
|
|
参考文献: |
|
|
[1] Pinz A. Object Categorization[J]. Foundations and Trends in Computer Graphics and Vision, 2005, 1(4): 255-353
[2] Dietterich T G, Lathrop R H, Lozano-Pérez T. Solving the Multiple Instance Problem with Axis-Parallel Rectangles[J]. Artificial Intelligence, 1997, 89(1/2): 31-71
[3] Zhou Z H, Zhang M L, Huang S J, et al. Multi-Instance Multi-Label Learning[J]. Artificial Intelligence, 2008, 176(1): 2291-2320
[4] 袁立明. 基于特征选择的嵌入空间多示例学习算法研究[D]. 哈尔滨: 哈尔滨工业大学, 2014:2-11 Yuan Liming. Embedding Space Based Multiple Instance Learning Algorithms with Feature Selection[D]. Harbin Institute of Technology, 2014: 2-11 (in Chinese)
[5] Amores J. Multiple Instance Classification: Review, Taxonomy and Comparative Study[J]. Artificial Intelligence, 2013, 201(4): 81-105
[6] Maron O, Lozano-Pérez T. A Framework for Multiple-Instance Learning[J]. Advances in Neural Information Processing Systems, 1998, 2000(2): 570-576
[7] Zhang Q, Goldman S A. EM-DD: An Improved Multiple-Instance Learning Technique[J]. Advances in Neural Information Processing Systems, 2002, 14: 1073-1080
[8] Andrews S, Hofmann T, Tsochantaridis I. Multiple Instance Learning with Generalized Support Vector Machines[C]//Eighteenth National Conference on Artificial Intelligence, 2002: 519-23
[9] Chen Y, Li J, Wang J Z. Categorization by Learning and Reasoning with Regions[M]. Springer US, 2004: 99-121
[10] Chen Y, Bi J, Wang J Z. MILES: Multiple-Instance Learning Via Embedded Instance Selection[J]. IEEE Trans on Pattern Analysis & Machine Intelligence, 2006, 28(12): 1931-47
[11] Schölkopf B, Platt J, Hofmann T. Efficient Sparse Coding Algorithms[C]//NIPS, 2007: 801-808
[12] 宋相法, 焦李成. 基于稀疏编码和集成学习的多示例多标记图像分类方法[J]. 电子与信息学报, 2013, (3): 622-626 Song Xiangfa, Jiao Licheng. A Multi-Instance Multi-Label Image Classification Method Based on Sparse Coding and Ensemble Learning[J]. Journal of electronics & information technology, 2013, (3): 622-626 (in Chinese)
[13] Yang J, Yu K, Gong Y, et al. Linear Spatial Pyramid Matching Using Sparse Coding for Image Classification[C]//2014: 1794-1801
[14] Foulds J, Frank E. Revisiting Multiple-Instance Learning Via Embedded Instance Selection[C]//2008: 300-310
[15] Fu Z, Robles-Kelly A, Zhou J. MILIS: Multiple Instance Learning with Instance Selection[J]. IEEE Trans on Pattern Analysis & Machine Intelligence, 2011, 33(5): 958-77
[16] Li W J, Yeung D Y. MILD: Multiple-Instance Learning via Disambiguation[J]. IEEE Trans on Knowledge & Data Engineering, 2009, 22(1): 76-89
[17] Tibshirani R. Regression Shrinkage and Selection via the Lasso[J]. J of the Royal Statistical Society, 1996, 58(1): 267-288
[18] Chang C C, Lin C J. LIBSVM: A Library for Support Vector Machines[J]. Acm Trans on Intelligent Systems & Technology, 2011, 2(3): 389-396
[19] Zhou Z H, Sun Y Y, Li Y F, Multi-Instance Learning by Treating Instances as Non-i.i.d. Samples[C]//International Conference on Machine Learning, 2009: 1249–1256 |
|
|
|
|
|
|
|
