|
|
论文:2019,Vol:37,Issue(3):465-470 |
|
|
引用本文： |
|
|
张鹏飞, 董敏周, 端军红. 基于集成GMM聚类的少标记样本图像分类[J]. 西北工业大学学报 |
|
|
ZHANG Pengfei, DONG Minzhou, DUAN Junhong. Classification of Few Labeled Images Based on Integrated GMM Clustering[J]. Northwestern polytechnical university |
|
|
|
|
|
|
|
| 基于集成GMM聚类的少标记样本图像分类 |
|
| 张鹏飞1, 董敏周1, 端军红2 |
|
1. 西北工业大学 航天学院, 陕西 西安 710072;
2. 空军工程大学 防空反导学院, 陕西 西安 710043 |
| 摘要: |
| 为了提高卷积神经网络训练的分类器分类准确率，往往需要大量的已标记数据，但有时已标记数据并不容易获得。针对少标记样本图像分类问题，提出基于集成GMM聚类与标签传递思想的解决方案，通过一定的规则给未标记数据赋予标签，将未标记数据转换成已标记数据用于模型的训练。在手写数字识别数据集上进行实验，结果表明新算法在少标记样本的情况下，结合集成GMM聚类的方法比只采用有标记样本训练得到的模型分类准确率有着较大提高，验证了该算法的有效性。 |
| 关键词:
集成GMM聚类
少标记样本
投票规则
|
|
| Classification of Few Labeled Images Based on Integrated GMM Clustering |
|
| ZHANG Pengfei1, DONG Minzhou1, DUAN Junhong2 |
|
1. School of Astronautics, Northwestern Polytechnical University, Xi'an 710072, China;
2. Air Defense Academy, Air Force Engineering University, Xi'an 710043, China |
| Abstract: |
| In order to improve the classifier classification accuracy of by using convolutional neural network training, a large amount of labeled data is often required, but sometimes labeled data is not easily obtained.This paper proposes a solution based on the idea of integrated GMM clustering and label delivery for classifying images with few labeled samples, assigning tags to unlabeled data through certain rules, and converting unlabeled data into labeled data for training of the model.In this paper, experiments are performed on hand-written digital recognition data sets. The results show that the present algorithm has a great improvement in the accuracy of model classification comparing with the method of using only labeled samples in the case of few labeled samples. The effectiveness of the present algorithm is validated. |
| Key words:
integrated GMM clustering
few labeled samples
voting rules
|
|
| 收稿日期: 2018-04-04
修回日期:
|
| DOI: 10.1051/jnwpu/20193730465 |
| 基金项目: 国家自然科学基金（11502300）资助 |
| 通讯作者:
Email： |
| 作者简介: 张鹏飞(1992-),西北工业大学硕士研究生,主要从事机器学习与图像处理研究。
|
|
|
|
|
|
|
|
参考文献: |
|
|
[1] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. ImageNet Classification with Deep Convolutional Neural Networks[J]. Advances in Neural Information Processing Systems, 2012, 25(2):1097-1105
[2] 罗帆,王厚峰. 结合RNN和CNN层次化网络的中文文本情感分类[J]. 北京大学学报, 2018, 54(3):459-465 LUO Fan, WANG Houfeng. The Chinese Text Sentiment Classification Combined with RNN and CNN Hierarchical Networks[J]. Journal of Peking University, 2018, 54(3):459-465(in Chinese)
[3] 钱程. 基于深度学习的人脸识别技术研究[D]. 成都:西南交通大学, 2017 QIAN Cheng. Research on Face Recognition Based on Deep Learning[D]. Chengdu, Southwest Jiaotong University, 2017(in Chinese)
[4] 戴礼荣,张仕良,黄智颖. 基于深度学习的语音识别技术现状与展望[J]. 数据采集与处理, 2017, 32(2):221-231 DAI Lirong, ZHANG Shiliang, HUANG Zhiying. The Present Situation and Prospect of Deep Recognition Based Speech Recognition Technology[J]. Data Acquisition and Processing, 2017, 32(2):221-231(in Chinese)
[5] 蔡毅,朱秀芳,孙章丽,等. 半监督集成学习综述[J]. 计算机科学, 2017, 44(增刊1):7-13 CAI Yi, ZHU Xiufang, SUN Zhangli, et al. An Overview of Semi-Supervised Integrated Learning[J]. Computer Science, 2017, 44(suppl 1):7-13(in Chinese)
[6] STAUFFER C, GRIMSON W E. Adaptive Background Mixture Models for Real-Time Tracking[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 1999:246-252
[7] 史倩玉. 复杂数据的集成聚类算法研究[D]. 太原:山西大学, 2017 SHI Qianyu. Research on Integrated Clustering Algorithm for Complex Data[D]. Taiyuan, Shanxi University, 2017(in Chinese)
[8] 鲁刚,余翔湛,张宏莉,等. 基于集成聚类的流量分类架构[J]. 软件学报, 2016, 27(11):2870-2883 LU Gang, YU Xiangzhan, ZHANG Hongli, et al. Flow Classification Architecture Based on Integrated Clustering[J]. Journal of Software, 2016, 27(11):2870-2883(in Chinese)
[9] 熊彪,江万寿,李乐林. 基于高斯混合模型的遥感影像半监督分类[J]. 武汉大学学报, 2011, 36(1):108-112 XIONG Biao, JIANG Wanshou, LI Lelin. Remote Sensing Image Semi-Supervised Classification Based on Gaussian Mixture Model[J]. Engineering Journal of Wuhan University, 2011, 36(1):108-112(in Chinese)
[10] 李航. 统计学习方法[M]. 北京:清华大学出版社, 2012 LI Hang. Statistical Learning Method[M]. Beijing, Tsinghua University Press, 2012(in Chinese) |
|
|
|
|
|
|
|
