SAR ATR中标签噪声不确定性建模与纠正  

Modeling and Correction of Label Noise Uncertainty for SAR ATR

作　　者：于跃 王琛 师君 陶重犇[1] 李良 唐欣欣 周黎明 韦顺军 张晓玲 YU Yue;WANG Chen;SHI Jun;TAO Chongben;LI Liang;TANG Xinxin;ZHOU Liming;WEI Shunjun;ZHANG Xiaoling(School of Electronic and Information Engineering,Suzhou University of Science and Technology,Suzhou 215000,China;School of Information and Communication Engineering,University of Electronic Science and Technology of China,Chengdu 611731,China;Sichuan Aerospace System Engineering Research Institute,Chengdu 610100,China;School of Information Science and Engineering,Chongqing Jiaotong University,Chongqing 400074,China;National Key Laboratory of Electromagnetic Space Security,Chengdu 610036,China)

机构地区：[1]苏州科技大学电子与信息工程学院,苏州215000 [2]电子科技大学信息与通信工程学院,成都611731 [3]四川航天系统工程研究所,成都610100 [4]重庆交通大学信息科学与工程学院,重庆400074 [5]电磁空间安全全国重点实验室,成都610036

出　　处：《雷达学报（中英文）》2024年第5期974-984,共11页Journal of Radars

基　　金：国家自然科学基金(62201375);江苏省自然科学基金(BK20220635);重庆市自然科学基金面上项目(CSTB2024NSCQMSX1762);重庆市教育委员会科学技术研究项目(KJQN202300756)。

摘　　要：深度监督学习在合成孔径雷达自动目标识别任务中的成功依赖于大量标签样本。但是,在大规模数据集中经常存在错误(噪声)标签,很大程度降低网络训练效果。该文提出一种基于损失曲线拟合的标签噪声不确定性建模和基于噪声不确定度的纠正方法:以损失曲线作为判别特征,应用无监督模糊聚类算法获得聚类中心和类别隶属度以建模各样本标签噪声不确定度;根据样本标签噪声不确定度将样本集划分为噪声标签样本集、正确标签样本集和模糊标签样本集,以加权训练损失方法分组处理训练集,指导分类网络训练实现纠正噪声标签。在MSTAR数据集上的实验证明,该文所提方法可处理数据集中混有不同比例标签噪声情况下的网络训练问题,有效纠正标签噪声。当训练数据集中标签噪声比例较小(40%)时,该文所提方法可纠正98.6%的标签噪声,并训练网络达到98.7%的分类精度。即使标签噪声比例很大(80%)时,该文方法仍可纠正87.8%的标签噪声,并训练网络达到82.3%的分类精度。The success of deep supervised learning in Synthetic Aperture Radar(SAR)Automatic Target Recognition(ATR)relies on a large number of labeled samples.However,label noise often exists in large-scale datasets,which highly influence network training.This study proposes loss curve fitting-based label noise uncertainty modeling and a noise uncertainty-based correction method.The loss curve is a discriminative feature to model label noise uncertainty using an unsupervised fuzzy clustering algorithm.Then,according to this uncertainty,the sample set is divided into different subsets:the noisy-label set,clean-label set,and fuzzylabel set,which are further used in training loss with different weights to correct label noise.Experiments on the Moving and Stationary Target Acquisition and Recognition(MSTAR)dataset prove that our method can deal with varying ratios of label noise during network training and correct label noise effectively.When the training dataset contains a small ratio of label noise(40%),the proposed method corrects 98.6%of these labels and trains the network with 98.7%classification accuracy.Even when the proportion of label noise is large(80%),the proposed method corrects 87.8%of label noise and trains the network with 82.3%classification accuracy.

关 键 词：合成孔径雷达 标签噪声 标签噪声纠正 标签噪声不确定性建模 模糊聚类算法 

分 类 号：TN957.52[电子电信—信号与信息处理]