PCA预训练的卷积神经网络目标识别算法

doi:10.3969/j.issn.1001-2400.2016.03.028

西安电子科技大学学报 ›› 2016, Vol. 43 ›› Issue (3): 161-166.doi: 10.3969/j.issn.1001-2400.2016.03.028

PCA预训练的卷积神经网络目标识别算法

史鹤欢;许悦雷;马时平;李岳云;李帅

(空军工程大学航空航天工程学院，陕西西安 710038)

收稿日期:2015-01-19 出版日期:2016-06-20 发布日期:2016-07-16
通讯作者: 史鹤欢
作者简介:史鹤欢(1990-)，男，空军工程大学博士研究生， E-mail: shihehuan1990@126.com．
基金资助:
国家自然科学基金资助项目(61379104，61372167)

Convolutional neural networks recognition algorithm based on PCA

SHI Hehuan;XU Yuelei;MA Shiping;LI Yueyun;LI Shuai

(Aeronautics and Astronautics Engineering College, Air Force Engineering Univ., Xi'an 710038, China)

Received:2015-01-19 Online:2016-06-20 Published:2016-07-16
Contact: SHI Hehuan

摘要/Abstract

摘要：

针对卷积神经网络对合成孔径雷达目标识别训练在标签数据不足，平移、旋转以及复杂情况下的识别率不高问题，提出一种优化的卷积神经网络目标识别算法．为克服标签数据不足，利用主成分分析非监督训练一组特征集初始化卷积神经网络；为提高训练速度，同时避免陷入过拟合，采用线性修正函数作为非线性函数；为增强鲁棒性，同时减小下采样对特征表示的影响，引入概率最大化下采样的方式，并在卷积层后对特征进行局部对比度标准化．实验表明，与传统的卷积神经网络相比，该算法对合成孔径雷达目标具有更高的识别率,并对图像各种形变以及复杂背景具有较好的鲁棒性．

关键词: 卷积神经网络, 主成分分析, 概率最大化下采样, 线性修正函数, 局部对比度标准化

Abstract:

To improve the insufficiency of Synthetic Aperture Radar(SAR) labeled training data for Convolutional Neural Networks(CNN) and the recognition rate for large variations, a novel CNN recognition algorithm is proposed. Firstly, a set of features is extracted from the original data by unsupervised training based on PCA as the initial filter set for CNN. Secondly, in order to accelerate the training speed while avoiding over-fitting, the Rectified Linear Units(ReLU) is adopted as the non-linear function. Thirdly, to strengthen robustness and mitigate the defects of pooling upon features, a probabilistic max-pooling sampling method is introduced and local contrast normalization is exploited on features after the convolutional layer. Experiments demonstrate that our algorithm outperforms the original CNN in recognition rate and achieves better robustness for large variations and complex background.

Key words: convolutional neural network, principal component analysis, probabilistic max-pooling, rectified linear units, local contrast normalization

中图分类号:

TP391.41

史鹤欢;许悦雷;马时平;李岳云;李帅. PCA预训练的卷积神经网络目标识别算法[J]. 西安电子科技大学学报, 2016, 43(3): 161-166.

SHI Hehuan;XU Yuelei;MA Shiping;LI Yueyun;LI Shuai. Convolutional neural networks recognition algorithm based on PCA[J]. Journal of Xidian University, 2016, 43(3): 161-166.

参考文献

［1］李帅, 许悦雷, 马时平, 等. 一种深度神经网络SAR遮挡目标识别方法［J］. 西安电子科技大学学报, 2015, 42(3): 154-160.
LI Shuai, XU Yuelei, MA Shiping, et al. New Method for SAR Occluded Targets Recognition Using DNN［J］. Journal of Xidian University, 2015, 42(3): 154-160.
［2］ KYIZHEUSKY A, SUTSKEVER I, HINTON G E, et al. Image Net Classification with Deep Convolutional Neural Networks［C］//Neural Information Processing Systems Conference. Nevada: Neural Information Processing Systems Foundation, 2012: 1106-1114.
［3］ DAN C, UELI M, JURGEN S. Multi-column Deep Neural Networks for Image Classification［C］//IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2012: 3642-3649.
［4］ KAVUKCUOGLU K, SERMANET P, BOUREAU Y, et al. Learning Convolutional Feature Hierarchies for Visual Recognition［C］//Advances in Neural Information Processing Systems Workshops. Real Hook: Neural Information Processing System Foundation, 2010: 1090-1098.
［5］ KAVUKCUOGLU K, RANZATO M, FERGUS R, et al. Learning Invariant Features through Topographic Filter Maps［C］//IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE Computer Society, 2009: 1605-1612.
［6］ ZEILER M D, FERGUS R. Visualizing and Understanding Convolutional Neural Networks［C］//European Conference on Computer Vision. Berlin: Springer International Publishing, 2014: 818-833.
［7］ LEE H, GROSSE R, RANGANATH R, et al. Convolutional Deep Belief Networks for Scalable Unsupervised Learning of Hierarchical Representation［C］//Proceedings of the 26th International Conference on Machine Learning. New York: ACM, 2009: 609-616.
［8］ BURGES C J C, PLATT J C, JANA S, et al. Distortion Discriminant Analysis for Audio Fingerprinting［J］. IEEE Transactions on Speech and Audio Processing, 2003, 11(3): 165-174.
［9］ LECUN Y, BOTTOU L, BENGIO Y. Gradient-based Learning Applied to Document Recognition［J］. Proceedings of the IEEE, 2008, 86 (11): 2278-2324.
［10］ NEUBAUER C. Evaluation of Convolutional Neural Networks for Visual Recognition［J］. IEEE Transactions on Neural Networks, 1998, 11(4): 685-696.
［11］ JARRETT K, KAVUKCUOGLU K, RANZATO M A, et al. What is the Best Multi-stage for Object Architecture?［C］//Proceedings of the IEEE 12th International Conference on Computer Vision. Piscataway: IEEE, 2009: 2146-2153.
［12］ BOUVRIE J. Notes on Convolutional Neural Networks ［EB/OL］. ［2014-11-24］. http://www.math.duke.edu/～jvb/papers/cnn_tutorial.pdf.

[1]	李源,崔玉爽,王伟. 一种基于字词双通道网络的文本情感分析方法[J]. 西安电子科技大学学报, 2021, 48(6): 179-186.
[2]	于浩洋,尹良,李书芳,吕顺. 生成对抗网络小样本雷达调制信号识别算法[J]. 西安电子科技大学学报, 2021, 48(6): 96-104.
[3]	董如婵,焦李成,赵进,沈维燕. 一种深度融合机制的遥感图像目标检测技术[J]. 西安电子科技大学学报, 2021, 48(5): 128-138.
[4]	孙彦景,魏力,张年龙,云霄,董锴文,葛敏,程小舟,侯晓峰. 联合DD-GAN和全局特征的井下人员重识别方法[J]. 西安电子科技大学学报, 2021, 48(5): 201-211.
[5]	周鹏,杨军. 采用神经网络架构搜索的遥感影像分割方法[J]. 西安电子科技大学学报, 2021, 48(5): 47-57.
[6]	杨云航,闵连权. 采用空洞卷积的多尺度融合草图识别模型[J]. 西安电子科技大学学报, 2021, 48(5): 92-99.
[7]	陈昌川,王海宁,黄炼,黄涛,李连杰,黄向康,代少升. 一种基于局部表征的面部表情识别算法[J]. 西安电子科技大学学报, 2021, 48(5): 100-109.
[8]	宋建锋,苗启广,王崇晓,徐浩,杨瑾. 注意力机制的多尺度单目标跟踪算法[J]. 西安电子科技大学学报, 2021, 48(5): 110-116.
[9]	张宇浩,程培涛,张书豪,王秀美. 一种自适应权重学习的轻量超分辨率重建网络[J]. 西安电子科技大学学报, 2021, 48(5): 15-22.
[10]	回海生,张雪英,吴泽林,李凤莲. 一种主辅路径注意力补偿的脑卒中病灶分割方法[J]. 西安电子科技大学学报, 2021, 48(4): 200-208.
[11]	王平,江雨泽,赵光辉. 目标检测的多尺度定位提升算法[J]. 西安电子科技大学学报, 2021, 48(3): 85-90.
[12]	郭泽坤,田隆,韩宁,王鹏辉,刘宏伟,陈渤. 采用CNN-SSD的雷达HRRP小样本目标识别方法[J]. 西安电子科技大学学报, 2021, 48(2): 7-14.
[13]	成磊,王玥,田春娜. 一种添加残差注意力机制的视觉目标跟踪算法[J]. 西安电子科技大学学报, 2020, 47(6): 148-157.
[14]	孔鑫,陈刚,龚国良,鲁华祥,毛文宇. 一种面向卷积神经网络加速器的高性能乘累加器[J]. 西安电子科技大学学报, 2020, 47(4): 55-63.
[15]	党吉圣,杨军. 多特征融合的三维模型识别与分割[J]. 西安电子科技大学学报, 2020, 47(4): 149-157.

PCA预训练的卷积神经网络目标识别算法

Convolutional neural networks recognition algorithm based on PCA

PDF (PC)

赞

可视化

被引次数

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 10