一种改进卷积神经网络的教学图像检索方法

doi:10.19665/j.issn1001-2400.2019.03.009

西安电子科技大学学报 ›› 2019, Vol. 46 ›› Issue (3): 52-58.doi: 10.19665/j.issn1001-2400.2019.03.009

一种改进卷积神经网络的教学图像检索方法

刘道华^1,²,崔玉爽¹,赵岩松¹,宋玉婷¹,王景慧¹

^1. 信阳师范学院计算机与信息技术学院,河南信阳 464000
^2. 信阳师范学院河南省教育大数据分析与应用重点实验室,河南信阳 464000

收稿日期:2018-11-21 出版日期:2019-06-20 发布日期:2019-06-19
作者简介:刘道华(1974-),男,教授,博士,E-mail: ldhzzx@163.com.
基金资助:
国家自然科学基金(31872704);河南省重点研发与推广专项(182102210537);河南省高等教育教学改革项目(2017SJGLX389);河南省教师教育课程改革项目(2019-JSJYYB-031)

Method for retrieving the teaching image based on the improved convolutional neural network

LIU Daohua^1,²,CUI Yushuang¹,ZHAO Yansong¹,SONG Yuting¹,WANG Jinghui¹

^1. School of Computer and Information Technology, Xinyang Normal Univ. , Xinyang 464000, China
^2. Henan Key Lab. of Analysis and Applications of Education Big Data, Xinyang Normal Univ. , Xinyang 464000, China

Received:2018-11-21 Online:2019-06-20 Published:2019-06-19

摘要/Abstract

摘要：

针对卷积神经网络在提取图像特征时所造成的特征信息损失以及降低高维度图像特征数据等问题,提出了一种改进卷积神经网络的图像检索优化方法。该方法首先利用融合的卷积层提取图像特征,并在融合的卷积层之间添加全连接层以减少特征信息的丢失;然后采用主成分分析法对高维的特征数据进行有效的降维处理;最后采用余弦相似度的方法进行相似度匹配,以实现相似图像的检索。采用当前经典的LeNet-L、LeNet-5等方法同文中方法在图像检索性能评价指标上进行对比实验。实验结果表明,所提出的检索方法比文中其他检索方法在查全率和平均查准率方面提高了3%27.3%。

关键词: 图像检索, 卷积神经网络, 特征提取, 主成分分析

Abstract:

Aiming at the loss of feature information caused by the convolution neural network in extracting image features and the reduction of high-dimensional image feature data, an image retrieval optimization scheme based on the improved convolution neural network is proposed. First, the image features are extracted through the convolutional layer of fusion, and a full connection layer is added between the convolutional layers of fusion to reduce the loss of image feature information. Then the PCA is used to effectively reduce the dimension of high-dimensional characteristic data and the high-dimensional feature vectors are mapped to the low-dimensional vector space. Finally, the cosine similarity method is used to match the similarity to achieve similar image retrieval. The proposed method is compared with classical methods such as LeNet-L and LeNet-5 in the performance of image retrieval. Experimental results show that the proposed retrieval method improves the recall rate and the average precision rate by at least 3%27.3% compared with classical methods.

Key words: image retrieval, convolution neural network, feature extraction, principal component analysis

中图分类号:

TP399

刘道华,崔玉爽,赵岩松,宋玉婷,王景慧. 一种改进卷积神经网络的教学图像检索方法[J]. 西安电子科技大学学报, 2019, 46(3): 52-58.

LIU Daohua,CUI Yushuang,ZHAO Yansong,SONG Yuting,WANG Jinghui. Method for retrieving the teaching image based on the improved convolutional neural network[J]. Journal of Xidian University, 2019, 46(3): 52-58.

图/表 5

图1

图2

图3

表2

图4

参考文献 11

[1]	BHUTE A N, MESHRAM B B . Text Based Approach for Indexing and Retrieval of Image and Video: a Review[J]. Advances in Vision Computing: An International Journal, 2014,1(1):27-38.
[2]	WAN J, WANG D, HOI S C H , et al. Deep Learning for Content-based Image Retrieval: a Comprehensive Study [C]//Proceedings of the 2014 ACM International Conference on Multimedia. New York: ACM, 2014: 157-166.
[3]	DEMIR B, BRUZZONE L . A Novel Active Learning Method in Relevance Feedback for Content-based Remote Sensing Image Retrieval[J]. IEEE Transactions on Geoscience and Remote Sensing, 2015,53(5):2323-2334. doi: 10.1109/TGRS.2014.2358804
[4]	许强, 李伟, 占荣辉 , 等. 一种改进的卷积神经网络SAR目标识别算法[J]. 西安电子科技大学学报, 2018,45(5):177-183.
	XU Qiang, LI Wei, ZHAN Ronghui , et al. Improved Algorithm for SAR Target Recognition Based on the Convolutional Neural Network[J]. Journal of Xidian University, 2018,45(5):177-183.
[5]	熊昌镇, 单艳梅, 郭芬红 . 结合主体检测的图像检索方法[J]. 光学精密工程, 2017,25(3):792-798.
	XIONG Changzhen, SHAN Yanmei, GUO Fenhong . Image Retrieval Method Based on Image Principal Part Detection[J]. Optics and Precision Engineering, 2017,25(3):792-798.
[6]	HE K M, ZHANG X Y, REN S Q , et al. Spatial Pyramid Pooling in Deep Convonlutional Networks for Visual Recogintion[J]. IEEE Transcations on Pattern Anlysis and Machine Intelligence, 2015,37(9):1904-1916. doi: 10.1109/TPAMI.2015.2389824
[7]	KRIZHEVSKY A, SUTSKEVER I, HINTON G E . ImageNet Classification with Deep Convolutional Neural Networks[J]. Communications of the ACM, 2017,60(6):84-90. doi: 10.1145/3098997
[8]	周飞燕, 金林鹏, 董军 . 卷积神经网络研究综述[J]. 计算机学报, 2017,40(6):1229-1251.
	ZHOU Feiyan, JIN Linpeng, DONG Jun . Review of Convolutional Neural Network[J]. Chinese Journal of Computers, 2017,40(6):1229-1251.
[9]	JIA Y, SHELHAMER E, DONAHUE J , et al. Caffe: Convolutional Architecture for Fast Feature Embedding [C]//Proceedings of the 2014 ACM International Conference on Multimedia. New York: ACM, 2014: 675-678.
[10]	刘海龙, 李宝安, 吕学强 , 等. 基于深度卷积神经网络的图像检索算法研究[J]. 计算机应用研究, 2017,34(12):3816-3819.
	LIU Hailong, LI Baoan, LV Xueqiang , et al. Image Retrieval Based on Deep Convolutional Neural Network[J]. Application Research of Computers, 2017,34(12):3816-3819.
[11]	TONG C, LI J, ZHU F M . A Convolutional Neural Network Based Method for Event Classification in Event-driven Multi-sensor Network[J]. Computers and Electrical Engineering, 2017,60(6):90-99. doi: 10.1016/j.compeleceng.2017.01.005

比较项	Top10	Top20	Top30	Top40	Top50
LeNet-5	0.781 3	0.721 6	0.686 3	0.650 3	0.638 5
Method	0.900 6	0.783 5	0.686 3	0.643 2	0.596 4
CNN	0.906 8	0.859 5	0.789 7	0.713 0	0.669 2
LeNet-L	0.908 9	0.862 5	0.839 5	0.819 3	0.799 7
文中方法	0.950 0	0.943 7	0.912 2	0.907 3	0.869 5

[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]	闫佳,曹玉东,任佳兴,陈冬昊,李晓会. 深度非对称压缩型哈希算法[J]. 西安电子科技大学学报, 2021, 48(5): 212-221.
[6]	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.
[7]	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.
[8]	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.
[9]	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.
[10]	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.
[11]	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.
[12]	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.
[13]	魏梓钰,杨曦,王楠楠,杨东,高新波. 互惠双向生成对抗网络用于跨模态行人重识别[J]. 西安电子科技大学学报, 2021, 48(2): 205-212.
[14]	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.
[15]	韩卓茜,王锋,陈沛,李卓伦. 一种历史特征辅助的模糊数据关联算法[J]. 西安电子科技大学学报, 2021, 48(2): 92-98.

一种改进卷积神经网络的教学图像检索方法

Method for retrieving the teaching image based on the improved convolutional neural network

RichHTML

PDF (PC)

赞

可视化

摘要/Abstract

引用本文

使用本文

图/表 5

参考文献 11

相关文章 15

Metrics

本文评价

推荐阅读 10