Saliency Detection Based on Unsupervised SDAE Network

LI Qing-wu; MA Yun-peng; ZHOU Ya-qin; XING Jun

doi:10.3969/j.issn.0372-2112.2019.04.015

您当前的位置：

首页 >

文章列表页 >

Saliency Detection Based on Unsupervised SDAE Network

更新时间：2025-07-16

- Saliency Detection Based on Unsupervised SDAE Network
- Acta Electronica Sinica Vol. 47, Issue 4, Pages: 871-879(2019)
- 作者机构：
  
  1. 河海大学物联网工程学院,江苏,常州,213022
  2. 常州市传感网与环境感知重点实验室,江苏,常州,213022
  3. 河海大学物联网工程学院,江苏,常州,213022
  4. 常州市传感网与环境感知重点实验室,江苏,常州,213022
- 作者简介：
- 基金信息：
  
  National Key Research and Development Program of China (No.2018YFC0406900);Key Research and Development Project of Jiangsu Province (No.BE2016071, No.BE2017057, No.BE2017648)
- DOI：10.3969/j.issn.0372-2112.2019.04.015
  CLC： TP751.1
- Published Online：25 April 2019，
  
  Published：2019
- 稿件说明：
移动端阅览
LI Qing-wu, MA Yun-peng, ZHOU Ya-qin, et al. Saliency Detection Based on Unsupervised SDAE Network[J]. Acta Electronica Sinica, 2019, 47(4): 871-879.
DOI：

LI Qing-wu, MA Yun-peng, ZHOU Ya-qin, et al. Saliency Detection Based on Unsupervised SDAE Network[J]. Acta Electronica Sinica, 2019, 47(4): 871-879. DOI： 10.3969/j.issn.0372-2112.2019.04.015.

摘要

针对现有的显著性检测算法检测目标类型单一、通用性差的问题，提出一种基于无监督栈式降噪自编码网络的显著性检测算法.该算法利用无监督栈式降噪自编码网络（Stacked Denoising Auto Encoder，SDAE）在多个尺度对原始图像进行稀疏重构，将原始图像与SDAE网络重构图像之间的差作为显著图，二值化后的显著图作为显著性目标检测结果.在SDAE网络训练过程中，将原始图像作为原始数据，网络重构的图像作为观察数据.为了提升网络训练效率，首先利用无监督逐层贪婪方法训练同结构的深度信念网络（Deep Belief Network，DBN），将训练得到的DBN网络参数设为SDAE网络的初始参数，再计算原始数据与观察数据之间的互信息作为网络收敛代价，利用反向传播进行网络参数微调.实验表明，该网络模型可以完成多类型目标的显著性检测，具有通用性好，准确度高等优点.

Abstract

The traditional saliency detection method is difficult to detect different kinds of saliency target simultaneously.In order to solve this problem

an algorithm based on unsupervised SDAE network is proposed in this paper.The stacked denoising auto-encoder (SDAE) network is used to sparsely reconstruct original image in multiple scales.The difference between the original image and the reconstructed image is used as a saliency map

and the binaryzation of the saliency map is used as salient detection result.In the process of SDAE network training

the original image is used as the original data and the reconstructed images are treated as observed data.In order to improve the efficiency of network training

the deep belief network (DBN) is trained by greedy method in each layer without supervising

and the network parameters are delivered to stacked denoising auto-encoder (SDAE) network as initial parameters.Then

the mutual information between the original data and the observed data is used as loss function

and the network parameters are tuned by backpropagation.The experiments show that the proposed algorithm can accomplish the saliency detection of various targets

which has the advantages of good universality and high accuracy.

关键词

Keywords

references

Views

390

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

A Spectral Algorithm Based on Weighted Sparse Coding for Visual Saliency Detection

Fusion Algorithm for Incomplete Fingerprint Orientation Field Reconstruction Based on Gradient Dispersion and Mutual Information

Genetic Sequential IB Algorithm

Novel Measures Based on Simple Schur Concave Function for Image Registration

An Adaptive Haar-Like MCTF Based Wavelet Video Coding Scheme

Related Author

YAO Xi-wen

CHENG Gong

HAN Jun-wei

GUO Lei

QIAN Xiao-liang

ZHANG Shao-le

SUN Song-lin

JING Xiao-jun

Related Institution

School of Automation, Northwestern Polytechnical University

School of Information and Communication Engineering, Beijing University of Posts and Telecommunications

School of Information EngineeringZhengzhou University.ZhengzhouHenan 450052China

Institute of Engineering TechnologyDongguan University of TechnologyDongguanGuangdong 523808China

School of Information Engineering,Zhengzhou University.

⁰