检索结果-南通市图书馆

A new progressive open-set recognition method with adaptive probability threshold

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2023年第11期36卷 297-310页

作者： zhunga liu Xuemeng HUI Yimin FU Department of Automation Northwestern Polytechnical UniversityXi’an 710072China

In the traditional pattern classification method,it usually assumes that the object to be classified must lie in one of given(known)classes of the training data ***,the training data set may not contain the class of some objects in practice,and this is considered as an Open-Set Recognition(OSR)*** this paper,we propose a new progressive open-set recognition method with adaptive probability *** the labeled training data and the test data(objects to be classified)are put into a common data set,and the k-Nearest Neighbors(k-NNs)of each object are sought in this common ***,we can determine the probability of object lying in the given *** the majority of k-NNs of the object are from labeled training data,this object quite likely belongs to one of the given classes,and the density of the object and its neighbors is taken into account ***,when most of k-NNs are from the unlabeled test data set,the class of object is considered very uncertain because the class of test data is unknown,and this object cannot be classified in this *** the objects belonging to known classes with high probability are all found,we re-calculate the probability of the other uncertain objects belonging to known classes based on the labeled training data and the objects marked with the estimated *** iteration will stop when the probabilities of all the objects belonging to known classes are not ***,a modified Otsu’s method is employed to adaptively seek the probability threshold for the final *** the probability of object belonging to known classes is smaller than this threshold,it will be assigned to the ignorant(unknown)class that is not included in training data *** other objects will be committed to a specific *** effectiveness of the proposed method has been validated using some experiments.

关键词： Data mining k-nearest neighbors Open-set recognition Object recognition The Otsu’s method

Evidential combination of augmented multi-source of information based on domain adaptation

在线全文

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2020年第11期63卷 38-55页

作者： Linqing HUANG zhunga liu Quan PAN Jean DEZERT School of Automation Northwestern Polytechnical University ONERA-The French Aerospace Lab

In the applications of domain adaptation(DA), there may exist multiple source domains, and each source domain usually provides some auxiliary information for object classification. The combination of such complementary knowledge from different source domains is helpful for improving the accuracy. We propose an evidential combination of augmented multi-source of information(ECAMI) method. The information sources are augmented at first by merging several randomly selected source domains to generate extra auxiliary information. We can obtain one piece of classification result with the assistance of each information source based on DA. Then these multiple classification results are combined by belief functions theory, which is expert at dealing with the uncertain information. Nevertheless, the classification results derived from different information sources may have different weights. The optimal weights are calculated by minimizing an given error criteria defined by the distance between the combination result and the ground truth using some training data. For each object, the augmented information sources will produce multiple classification results that will be discounted by the learnt weights under the belief functions framework. Then the combination of these discounted results is employed to make the final class decision. The effectiveness of ECAMI is evaluated with respect to some related methods based on several real data sets, and the experimental results show that ECAMI can significantly improve the classification accuracy.

关键词： information fusion domain adaptation evidence theory belief functions pattern classification

Conditional self-attention generative adversarial network with differential evolution algorithm for imbalanced data classification

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2023年第3期36卷 303-315页

作者： Jiawei NIU zhunga liu Quan PAN Yanbo YANG Yang LI Department of Automation Northwestern Polytechnical UniversityXi’an 710072China

Imbalanced data classification is an important research topic in real-world applications,like fault diagnosis in an aircraft manufacturing *** over-sampling method is often used to solve this *** generates samples according to the distance between minority ***,the traditional over-sampling method may change the original data distribution,which is harmful to the classification *** this paper,we propose a new method called Conditional SelfAttention Generative Adversarial Network with Differential Evolution(CSAGAN-DE)for imbalanced data *** new method aims at improving the classification performance of minority data by enhancing the quality of the generation of minority *** CSAGAN-DE,the minority data are fed into the self-attention generative adversarial network to approximate the data distribution and create new data for the minority ***,the differential evolution algorithm is employed to automatically determine the number of generated minority data for achieving a satisfactory classification *** experiments are conducted to evaluate the performance of the new CSAGAN-DE *** results show that the new method can efficiently improve the classification performance compared with other related methods.

关键词： Classification Generative adversarial network Imbalanced data Optimization Over-sampling

Combination of classifiers with incomplete frames of discernment

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2022年第5期35卷 145-157页

作者： zhunga liu Jingfei DUAN Linqing HUANG Jean DEZERT Yongqiang ZHAO School of Automation Northwestern Polytechnical UniversityXi’an 710072China ONERA–The French Aerospace Lab Palaiseau 91761France

The methods for combining multiple classifiers based on belief functions require to work with a common and complete(closed)Frame of Discernment(Fo D)on which the belief functions are defined before making their *** theoretical requirement is however difficult to satisfy in practice because some abnormal(or unknown)objects that do not belong to any predefined class of the Fo D can appear in real classification *** classifiers learnt using different attributes information can provide complementary knowledge which is very useful for making the classification but they are usually based on different Fo *** order to clearly identify the specific class of the abnormal objects,we propose a new method for combination of classifiers working with incomplete frames of discernment,named CCIF for *** is a progressive detection method that select and add the detected abnormal objects to the training data *** one pattern can be considered as an abnormal object by one classifier and be committed to a specific class by another one,a weighted evidence combination method is proposed to fuse the classification results of multiple *** new method offers the advantage to make a refined classification of abnormal objects,and to improve the classification accuracy thanks to the complementarity of the *** experimental results are given to validate the effectiveness of the proposed method using real data sets.

关键词： Abnormal object Belief functions Classifier fusion Evidence theory Detection

Special Topic: Information Fusion

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2022年第5期 3-3页

作者： zhunga liu Wei XIONG School of Automation Northwestern Polytechnical University Information Fusion Research Institute Naval Aviation University

Information fusion technique plays an important role in target detection, tracking and recognition. It may contain some uncertain information in sensor observation because of the influence of noisy environment. We can efficiently reduce the uncertainty and improve the robustness of systems via multi-source information fusion thanks to the complementary knowledge from different sources. In applications, the sensing environment becomes more and more complex, and new kinds of sensors are emerging. This causes the heterogeneous and even conflicting information. So information fusion is faced with great challenges. Recently, many new fusion mechanisms have been developed, e.g.,

关键词：

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

证据冲突下自适应融合目标识别算法

航空学报 2010年第7期31卷 1426-1432页

作者：刘准钆程咏梅潘泉苗壮西北工业大学自动化学院陕西西安710072

利用证据理论对空中目标识别系统的观测信息融合时,Dempster规则对低冲突信息的融合结果较为理想,但无法对高冲突信息有效融合。Dubois&Prade(DP)规则及证据折扣法可对高冲突信息进行合理融合。为使不同融合方法发挥各自优势,提出一种... 详细信息

利用证据理论对空中目标识别系统的观测信息融合时,Dempster规则对低冲突信息的融合结果较为理想,但无法对高冲突信息有效融合。Dubois&Prade(DP)规则及证据折扣法可对高冲突信息进行合理融合。为使不同融合方法发挥各自优势,提出一种自适应融合算法。首先将矛盾因子和证据距离两者结合以更全面地表示证据冲突程度,当冲突较小时,选用Dempster规则,反之,根据冲突的具体情况选择使用DP规则或证据折扣法。通过目标识别实验对多种算法进行了对比,表明本文算法既能对高冲突证据进行合理融合,又能使融合结果快速收敛,可以有效地提高识别速度及正确率。

关键词：信息融合目标识别证据理论矛盾因子证据距离

卷积神经网络在雷达自动目标识别中的研究进展

在线全文

学校读者我要写书评

暂无评论

电子与信息学报 2020年第1期42卷 119-131页

作者：贺丰收何友刘准钆徐从安西北工业大学自动化学院西安710129 航空工业雷华电子技术研究所射频综合仿真实验室无锡214063 海军航空大学信息融合研究所烟台264001

自动目标识别(ATR)是雷达信息处理领域的重要研究方向。由于卷积神经网络(CNN)无需进行特征工程,图像分类性能优越,因此在雷达自动目标识别领域研究中受到越来越多的关注。该文综合论述了CNN在雷达图像处理中的应用进展。首先介绍了雷... 详细信息

自动目标识别(ATR)是雷达信息处理领域的重要研究方向。由于卷积神经网络(CNN)无需进行特征工程,图像分类性能优越,因此在雷达自动目标识别领域研究中受到越来越多的关注。该文综合论述了CNN在雷达图像处理中的应用进展。首先介绍了雷达自动目标识别相关知识,包括雷达图像的特性,并指出了传统的雷达自动目标识别方法局限性。给出了CNN卷积神经网络原理、组成和在计算机视觉领域的发展历程。然后着重介绍了CNN在雷达自动目标识别中的研究现状,其中详细介绍了合成孔径雷达(SAR)图像目标的检测与识别方法。接下来对雷达自动目标识别面临的挑战进行了深入分析。最后对CNN新理论、新模型,以及雷达新成像技术和未来复杂环境下的应用进行了展望。

关键词：自动目标识别目标检测合成孔径雷达卷积神经网络

基于三级信息融合结构的多平台多雷达目标识别算法

在线全文

学校读者我要写书评

暂无评论

西北工业大学学报 2012年第3期30卷 367-372页

作者：吴俊程咏梅曲圣杰潘泉刘准钆西北工业大学自动化学院陕西西安710072

为了解决多平台多雷达目标识别时空信息融合过程中存在的时效性不高,识别特征信息模糊和雷达识别结果高冲突等问题,提出了一种基于三级信息融合结构的多平台多雷达目标识别算法。一级特征融合中,基于模糊评判法解决了多特征识别信息模... 详细信息

为了解决多平台多雷达目标识别时空信息融合过程中存在的时效性不高,识别特征信息模糊和雷达识别结果高冲突等问题,提出了一种基于三级信息融合结构的多平台多雷达目标识别算法。一级特征融合中,基于模糊评判法解决了多特征识别信息模糊问题。二级空域融合中,基于DS融合规则和折扣融合规则,提出了以平均证据距离为阈值的自适应证据融合算法。三级时域融合基于多雷达序列信息,给出了时域自适应融合算法。仿真计算结果表明文中提出的融合结构能够在保证结果正确的同时,每个周期都能输出结果,时效性较之以往的融合结构有了明显改善。文中的自适应证据融合算法能够克服证据间可能存在的高冲突性,并且与采用折扣证据融合算法相比计算量显著减少。

关键词：自适应证据融合算法模糊综合评判目标识别三级信息融合结构

在线全文

学校读者我要写书评

暂无评论

基于证据理论的水声多目标优选方法

西北工业大学学报 2014年第3期32卷 429-433页

作者：禹亮程咏梅陈克喆刘建新刘准钆西北工业大学自动化学院陕西西安710072

针对水声复杂对抗环境下,诱饵与潜艇并存较难通过频谱分析进行辨识的问题,提出了在跟踪过程中,结合目标声学特征与运动特征信息来对目标进行优选并通过证据理论实现该方法。通过对水声目标属性的研究给出了优选所需的相关属性,对各属性... 详细信息

针对水声复杂对抗环境下,诱饵与潜艇并存较难通过频谱分析进行辨识的问题,提出了在跟踪过程中,结合目标声学特征与运动特征信息来对目标进行优选并通过证据理论实现该方法。通过对水声目标属性的研究给出了优选所需的相关属性,对各属性特征进行建模,提出了各属性的隶属度的计算方法,分别进行声学属性和运动属性的融合,最后进行声学属性与运动属性的加权融合实现算法。离散仿真和按照对抗策略动态仿真实验结果表明,该方法可有效对水声目标进行优选。

关键词：声学目标识别信息融合声学特征运动特征证据理论