A Novel Semi-Supervised Learning Method Based on Fast Search and Density Peaks

A Novel Semi-Supervised Learning Method Based on Fast Search and Density Peaks
复制标题

一种基于快速搜索和密度峰值的新型半监督学习方法

DOI:
10.1155/2019/6876173
复制
发表时间:
2019-02
期刊:
影响因子:
2.3
通讯作者:
Zhou Huiyu
Zhou Huiyu
中科院分区:
工程技术4区
文献类型:
--
作者:
Gao Fei;Huang Teng;Sun Jinping;Hussain Amir;Yang Erfu;Zhou Huiyu

文献摘要

参考文献

被引文献

相似文献

雷达图像识别是遥感领域的一个研究热点。在充分标注样本的条件下,识别算法可以取得良好的分类效果。然而,标记的样品是稀缺的,获得成本高。本文的主要目的是在有限的标记样本的情况下,如何利用这些未标记的样本来提高识别算法的性能。这是一个半监督学习问题。然而,与现有的半监督学习方法不同,我们不直接使用未标记的样本,而是在使用它们之前寻找安全可靠的未标记样本。本文提出了两种新的半监督学习方法:基于快速搜索和密度峰值的半监督学习方法(S2DP)和迭代S2DP方法(IS2DP)。当标记样本满足一定要求时,S2DP基于加权核Fisher判别分析(WKFDA),采用快速搜索和密度峰聚类方法检测出可靠的未标记样本。然后,设计了一种基于聚类信息的标记方法(LCI)来标记未标记的样本。当标记样本不足时,使用IS2DP迭代搜索可靠的未标记样本进行半监督。然后,将这些样本添加到标记样本中,以提高S2DP的识别性能。在实验中,使用真实的雷达图像来验证我们提出的算法在处理标记样本稀缺性方面的性能。此外,我们的算法与几种具有类似结构的半监督深度学习方法进行了比较。实验结果表明,该算法具有更好的稳定性比这些方法。
Radar image recognition is a hotspot in the field of remote sensing. Under the condition of sufficiently labeled samples, recognition algorithms can achieve good classification results. However, labeled samples are scarce and costly to obtain. Our major interest in this paper is how to use these unlabeled samples to improve the performance of a recognition algorithm in the case of limited labeled samples. This is a semi-supervised learning problem. However, unlike the existing semi-supervised learning methods, we do not use unlabeled samples directly and, instead, look for safe and reliable unlabeled samples before using them. In this paper, two new semi-supervised learning methods are proposed: a semi-supervised learning method based on fast search and density peaks (S2DP) and an iterative S2DP method (IS2DP). When the labeled samples satisfy a certain requirement, S2DP uses fast search and a density peak clustering method to detect reliable unlabeled samples based on the weighted kernel Fisher discriminant analysis (WKFDA). Then, a labeling method based on clustering information (LCI) is designed to label the unlabeled samples. When the labeled samples are insufficient, IS2DP is used to iteratively search for reliable unlabeled samples for semi-supervision. Then, these samples are added to the labeled samples to improve the recognition performance of S2DP. In the experiments, real radar images are used to verify the performance of our proposed algorithm in dealing with the scarcity of the labeled samples. In addition, our algorithm is compared against several semi-supervised deep learning methods with similar structures. Experimental results demonstrate that the proposed algorithm has better stability than these methods.
高分三号任务多模式SAR原始数据模拟与并行加速
DOI: 10.1109/jstars.2017.2787728
发表时间: 2018-01
影响因子: 5.5
作者:
Zhang Fan;Yao Xiaojie;Tang Hanyuan;Yin Qiang;Hu Yuxin;Lei Bin
通讯作者: Lei Bin
DOI: 10.1007/978-1-4899-7687-1_810
发表时间: 2017
期刊: --
影响因子: --
作者:
Xinhua Zhang
通讯作者: Xinhua Zhang
DOI: 10.1201/9781003139041-11
发表时间: 2021-03
期刊: An Introduction to IoT Analytics
影响因子: --
作者:
Harry G. Perros
通讯作者: Harry G. Perros
DOI: 10.1007/978-1-4471-5571-3_16
发表时间: 2014
期刊: Fuel
影响因子: 7.4
作者:
Ke-Lin Du;M. Swamy
通讯作者: Ke-Lin Du;M. Swamy
使用基于 QPSO 的多重 KFDA 信号处理对电子鼻信号进行特征提取(开放获取)
DOI: 10.3390/s18020388
发表时间: 2018-01-29
期刊: Sensors (Basel, Switzerland)
影响因子: --
作者:
Wen T;Yan J;Huang D;Lu K;Deng C;Zeng T;Yu S;He Z
通讯作者: He Z