基于无监督栈式降噪自编码网络的显著性检测算法

doi:10.3969/j.issn.0372-2112.2019.04.015

电子学报 ›› 2019, Vol. 47 ›› Issue (4): 871-879.DOI: 10.3969/j.issn.0372-2112.2019.04.015

基于无监督栈式降噪自编码网络的显著性检测算法

李庆武^1,2, 马云鹏^1,2, 周亚琴^1,2, 邢俊^1,2

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

收稿日期:2018-01-29 修回日期:2018-08-01 出版日期:2019-04-25
- 作者简介:
- 李庆武 男,1964年12月出生,河南新乡人.分别于郑州大学获得学士学位、西安电子科技大学大学获得硕士学位、河海大学获得博士学位.现为河海大学物联网工程学院副院长、教授、博士生导师.主要研究方向为智能视觉感知、信息获取与智能系统.E-mail:li_qingwu@163.com;马云鹏 男,1993年8月出生,山东-城人.2015年于河海大学获得学士学位,现为河海大学博士研究生.主要研究方向为数字图像处理、智能视觉感知.E-mail:yunpengma_hhu@163.com
- 基金资助:
- 国家重点研发计划 (No.2018YFC0406900）; 江苏省重点研发计划 (No.BE2016071，No.BE2017057，No.BE2017648）

Saliency Detection Based on Unsupervised SDAE Network

LI Qing-wu^1,2, MA Yun-peng^1,2, ZHOU Ya-qin^1,2, XING Jun^1,2

1. College of Internet of Things Engineering, Hohai University, Changzhou, Jiangsu 213022, China;
2. Changzhou Key Laboratory of Sensor Networks and Environmental Sensing, Changzhou, Jiangsu 213022, China

Received:2018-01-29 Revised:2018-08-01 Online:2019-04-25 Published:2019-04-25
- Supported by:
- National Key Research and Development Program of China (No.2018YFC0406900); Key Research and Development Project of Jiangsu Province (No.BE2016071, No.BE2017057, No.BE2017648)

摘要/Abstract

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.

关键词: 显著性检测, 无监督网络, 栈式降噪自编码, 深度信念网络, 互信息

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.

Key words: saliency detection, unsupervised network, stacked denoising auto-encoder (SDAE), deep belief network (DBN), mutual information

中图分类号:

TP751.1

李庆武, 马云鹏, 周亚琴, 等. 基于无监督栈式降噪自编码网络的显著性检测算法[J]. 电子学报, 2019, 47(4): 871-879.

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.

参考文献

[1] REN Z,GAO S,CHIA L T,et al.Region-based saliency detection and its application in object recognition[J].IEEE Transactions on Circuits & Systems for Video Technology,2014,24(5):769-779.
[2] 丁鹏,张叶,贾平,等.基于视觉显著性的海面舰船检测技术[J].电子学报,2018,46(1):127-134. DING Peng,ZHANG Ye,JIA Ping,et al.Ship detection on sea surface based on visual saliency[J].Acta Electronica Sinica,2018,46(1):127-134.(in Chinese)
[3] YANG X,QIAN X,XUE Y.Scalable mobile image retrieval by exploring contextual saliency[J].IEEE Transactions on Image Processing,2015,24(6):1709-1721.
[4] 冯松鹤,郎丛妍,须德.一种融合图学习与区域显著性分析的图像检索算法[J].电子学报,2011,39(10):2288-2294. FENG Song-he,LANG Cong-yan,XU De.Combining graph learning and region saliency analysis for content-based image retrieval[J].Acta Electronica Sinica,2011,39(10):2288-2294.(in Chinese)
[5] 姜维,卢朝阳,李静,等.基于视觉显著性与文字置信图的场景文字的背景抑制方法[J].电子学报,2015,43(9):62-68. JIANG Wei,LU Zhao-yang,LI Jing,et al.Visual saliency and text confidence map based background suppression for scene text[J].Acta Electronica Sinica,2015,43(9):62-68.(in Chinese)
[6] 宋腾飞,刘政怡.中心矩形构图先验的显著目标检测[J].中国图象图形学报,2017,22(3):315-326. SONG T-F,LIU Z-Y.Saliency detection based on center rectangle composition prior[J].Journal of Image and Graphics,2017,22(3):315-326.(in Chinese)
[7] ZHU W,LIANG S,WEI Y,et al.Saliency optimization from robust background detection[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2014.2814-282.
[8] LI X,LU H,ZHANG L,et al.Saliency detection via dense and sparse reconstruction[A].IEEE International Conference on Computer Vision[C].US:IEEE ICCV,2013.2976-2983.
[9] MA Y F,ZHANG H J.Contrast-based image attention analysis by using fuzzy growing[A].Eleventh ACM International Conference on Multimedia[C].US:ACM,2003.374-381.
[10] CHENG M M,ZHANG G X,MITRA N J,et al.Global contrast based salient region detection[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2011.409-416.
[11] HUANG F,JINQING Q,LU H,et al.Salient object detection via multiple instance learning[J].IEEE Transactions on Image Processing,2017,26(4):1911-1922.
[12] 刘峰,沈同圣,娄树理,等.全局模型和局部优化的深度网络显著性检测[J].光学学报,2017,37(12):1215005. LIU F,SHEN T-S,LOU S-L,et al.Deep saliency detection based on global model and local optimization[J].Acta Optica Sinica,2017,37(12):1215005.(in Chinese)
[13] HINTON G E,OSINDERO S,TEH Y W.A fast learning algorithm for deep belief nets[J].Neural Computation,2006,18(7):1527.
[14] ROUX N L,BENGIO Y.Representational power of restricted boltzmann machines and deep belief networks[J].Neural Computation,2008,20(6):1631-1649.
[15] SAINATH T N,KINGSBURY B,RAMABHADRAN B.Auto-encoder bottleneck features using deep belief networks[A].IEEE International Conference on Acoustics,Speech and Signal Processing[C].US:IEEE,2012.4153-4156.
[16] JUDD T,EHINGER K,DURAND F,et al.Learning to predict where humans look[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2009.2106-2113.
[17] BRUCE N D B,TSOTSOS J K.Saliency based on information maximization[A].International Conference on Neural Information Processing Systems[C].US:MIT Press 2005.155-162.
[18] BORJI A,ITTI L.State-of-the-art in visual attention modeling[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2013,35(1):185.
[19] RICHE N,MANCAS M,DUVINAGE M,et al.RARE2012:A multi-scale rarity-based saliency detection with its comparative statistical analysis[J].Signal Processing Image Communication,2013,28(6):642-658.
[20] HAREL J,KOCH C,PERONA P.Graph-based visual saliency[J].Advances in Neural Information Processing Systems,2007,19:545-552.
[21] BRUCE N,TSOTSOS J.Attention based on information maximization[J].Journal of Vision,2007,7(9):950-950.
[22] DUAN L,WU C,MIAO J,et al.Visual saliency detection by spatially weighted dissimilarity[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2011.473-480.
[23] BORJI A,ITTI L.Exploiting local and global patch rarities for saliency detection[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2012.478-485.
[24] HOU X,HAREL J,KOCH C.Image signature:highlighting sparse salient regions[J].IEEE Transactions on Pattern Analysis & Machine Intelligence,2012,34(1):194.
[25] VIKRAM T N,TSCHEREPANOW M,WREDE B.A saliency map based on sampling an image into random rectangular regions of interest[J].Pattern Recognition,2012,45(9):3114-3124.
[26] EEDEM E,ERDEM A.Visual saliency estimation by nonlinearly integrating features using region covariances[J].Journal of Vision,2013,13(4):11.
[27] FANG S,LI J,TIAN Y,et al.Learning discriminative subspaces on random contrasts for image saliency analysis[J].IEEE Transactions on Neural Networks & Learning Systems,2017,28(5):1095.
[28] TAVAKOLI H R,LAAKSONEN J.Bottom-up fixation prediction using unsupervised hierarchical models[A].Computer Vision ACCV 2016 Workshops[C].Berlin Heidelberg:Springer,2017.287-302.
[29] ZHAI Y,SHAH M.Visual attention detection in video sequences using spatiotemporal cues[A].ACM International Conference on Multimedia[C].US:ACM,2006.815-824.
[30] ACHANTA R,ESTRADA F,WILS P,et al.Salient region detection and segmentation[A].Proc ICVS[C].2008,5008:66-75.
[31] ACHANTA R,HEMAMI S,ESTRADA F,et al.Frequency-tuned salient region detection[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2009.1597-1604.
[32] GOFERMAN S,ZELNIK-MANOR L,TAL A.Context-aware saliency detection[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2010.2376-2383.
[33] RAHTU E,KANNALA J,SALO M,et al.Segmenting salient objects from images and videos[A].Computer Vision,ECCV 2010[C].Berlin Heidelberg:Springer,2010.366-379.
[34] Ran M,Tal A,ZELNIK-MANOR L.What makes a patch distinct?[A].Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition[C].US:IEEE CVPR,2013.1139-1146.
[35] YANG C,ZHANG L,LU H.Graph-regularized saliency detection with convex-hull-based center prior[J].IEEE Signal Processing Letters,2013,20(7):637-640.
[36] LI H,LU H,LIN Z,et al.Inner and inter label propagation:salient object detection in the wild[J].IEEE Transactions on Image Processing,2015,24(10):3176-3186.

基于无监督栈式降噪自编码网络的显著性检测算法

Saliency Detection Based on Unsupervised SDAE Network

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	陈星宇, 叶锋, 黄添强, 翁彬, 陈家祯, 林文忠. 融合小型深度生成模型的显著性检测[J]. 电子学报, 2021, 49(4): 768-774.
[2]	罗映雪, 贾博, 裘旭益, 邓平煜, 任和, 吴奇. 基于Gamma深度信念网络的飞行员脑疲劳状态识别[J]. 电子学报, 2020, 48(6): 1062-1070.
[3]	尤殿龙, 郭松, 赵春慧, 原福永, 申利民, 陈真. 面向分类的流特征在线特征选择算法[J]. 电子学报, 2020, 48(2): 321-332.
[4]	丁颖, 刘延伟, 刘金霞, 刘科栋, 王利明, 徐震. 虚拟现实全景图像显著性检测研究进展综述[J]. 电子学报, 2019, 47(7): 1575-1583.
[5]	彭丝雨, 周到, 张家琦, 王宇, 高军峰. 基于互信息的脑网络及测谎研究[J]. 电子学报, 2019, 47(7): 1551-1556.
[6]	杨小冈, 李维鹏, 马玛双. 概率框架下多特征显著性检测算法[J]. 电子学报, 2019, 47(11): 2378-2385.
[7]	张峰, 钟宝江. 基于兴趣目标的图像检索[J]. 电子学报, 2018, 46(8): 1915-1923.
[8]	李梦, 刘星, 任泽民. 联合Sobolev梯度的相场变分模型应用于图像显著性检测[J]. 电子学报, 2018, 46(7): 1683-1690.
[9]	杨兴明, 王雨廷, 谢昭, 吴克伟. 基于“轮廓-区域”多层互补特性的显著性检测[J]. 电子学报, 2018, 46(11): 2688-2695.
[10]	雷维嘉, 宋海娜, 陈胜男, 谢显中. 基于无率纠错码和高阶QAM调制的收端速率自适应方案[J]. 电子学报, 2017, 45(8): 2001-2007.
[11]	马润年, 王班, 王刚, 郭晓成, 刘文斌. 基于互信息的通信网络节点重要性度量方法[J]. 电子学报, 2017, 45(3): 747-752.
[12]	张晴晖, 张榆锋, 林文晶, 周屹, 章克信, 高莲. 颈动脉B超估计血管壁搏动位移的干扰抑制[J]. 电子学报, 2017, 45(12): 2903-2908.
[13]	续欣莹, 张扩, 谢珺, 谢刚. 基于互信息下粒子群优化的属性约简算法[J]. 电子学报, 2017, 45(11): 2695-2704.
[14]	宋岚, 文堂柳, 黎海生, 王杉. 基于云模型、图论和互信息的遥感影像分割方法[J]. 电子学报, 2015, 43(8): 1518-1525.
[15]	黄虹, 张建秋. 一个图像质量盲评估的统计测度[J]. 电子学报, 2014, 42(7): 1419-1423.