Image Retrieval Based on Convolutional Neural Network and Kernel-Based Supervised Hashing

KE Sheng-cai; ZHAO Yong-wei; LI Bi-cheng; PENG Tian-qiang

doi:10.3969/j.issn.0372-2112.2017.01.022

您当前的位置：

首页 >

文章列表页 >

Image Retrieval Based on Convolutional Neural Network and Kernel-Based Supervised Hashing

更新时间：2025-07-16

- Image Retrieval Based on Convolutional Neural Network and Kernel-Based Supervised Hashing
- Acta Electronica Sinica Vol. 45, Issue 1, Pages: 157-163(2017)
- 作者机构：
  
  1. 解放军信息工程大学信息系统工程学院,河南,郑州,450001
  2. 武警工程大学电子技术系,陕西,西安,710000
  3. 河南工程学院计算机学院,河南,郑州,450001
  4. 解放军信息工程大学信息系统工程学院,河南,郑州,450001
  5. 武警工程大学电子技术系,陕西,西安,710000
  6. 河南工程学院计算机学院,河南,郑州,450001
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China (No.60872142, No.61301232)
- DOI：10.3969/j.issn.0372-2112.2017.01.022
  CLC： TP391
- Published Online：25 January 2017，
  
  Published：2017
- 稿件说明：
移动端阅览
KE Sheng-cai, ZHAO Yong-wei, LI Bi-cheng, et al. Image Retrieval Based on Convolutional Neural Network and Kernel-Based Supervised Hashing[J]. Acta Electronica Sinica, 2017, 45(1): 157-163.
DOI：

KE Sheng-cai, ZHAO Yong-wei, LI Bi-cheng, et al. Image Retrieval Based on Convolutional Neural Network and Kernel-Based Supervised Hashing[J]. Acta Electronica Sinica, 2017, 45(1): 157-163. DOI： 10.3969/j.issn.0372-2112.2017.01.022.

摘要

当前主流的图像检索方法采用的视觉特征，缺乏自主学习能力，导致其图像表达能力不强，此外，传统的特征索引方法检索效率较低，难以适用于大规模图像数据.针对这些问题，本文提出了一种基于卷积神经网络和监督核哈希的图像检索方法.首先，利用卷积神经网络的学习能力挖掘训练图像内容的内在隐含关系，提取图像深层特征，增强特征的视觉表达能力和区分性；然后，利用监督核哈希方法对高维图像深层特征进行监督学习，并将高维特征映射到低维汉明空间中，生成紧致的哈希码；最后，在低维汉明空间中完成对大规模图像数据的有效检索.在ImageNet-1000和Caltech-256数据集上的实验结果表明，本文方法能够有效地增强图像特征的表达能力，提高图像检索效率，优于当前主流方法.

Abstract

The visual features of the state-of-the-art image retrieval methods lack of learning ability

which lead to low expression ability.And the efficiency of traditional index methods is fairly low for large image database.In view of this

an image retrieval method based on convolutional neural network and kernel-based supervised Hashing is proposed.Firstly

a large convolutional neural network is employed to learn the intrinsic implications of training images so as to improve the distinguish ability and expression ability of visual feature.Secondly

kernel-based supervised Hashing is applied to learn from the high-dimensional visual feature and map into low-dimensional hamming space and achieve compact Hash codes.Finally

image retrieval is accomplished in low-dimensional hamming space.Experimental results of ImageNet-1000 and Caltech-256 datasets indicate that the expression ability of visual feature is effectively improved and the image retrieval performance is substantially boosted compared with the state-of-the-art methods.

关键词

Keywords

references

Views

1464

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Operator Fusion Method and Hardware Architecture Design Based on Non-Standard Operators

MalMKNet: A Multi-Scale Convolutional Neural Network Used for Malware Classification

Research of Lightweight Convolution Neural Network Based on Feature Expansion Convolution

Mobile_BLNet: Optimization Design of Lightweight Convolutional Neural Network Based on Big-Little Net

Image Super-Resolution Reconstruction Based on Lightweight Multi-Scale Channel Attention Network

Related Author

WANG Ying

GAO Lan

ZHANG Zhe

LIU Xin

WU Yi-xiong

ZHANG Wei-gong

ZHANG Dan-dan

SONG Ya-fei

Related Institution

College of Information Engineering, Capital Normal University

School of Mathematical Science, Capital Normal University

Faculty of Software Technologics, Shanxi Agricultural University

Institute of Air Defense and Anti-missile， Air Force Engineering University

School of Computer Science and Technology, China University of Mining and Technology

⁰