您的位置:山东大学 -> 科技期刊社 -> 《山东大学学报(工学版)》

山东大学学报 (工学版) ›› 2021, Vol. 51 ›› Issue (4): 1-7.doi: 10.6040/j.issn.1672-3961.0.2020.266

• 机器学习与数据挖掘 •    下一篇

图像依赖的显著图融合方法

梁晔1,2(),马楠2,刘宏哲1   

  1. 1. 北京联合大学信息服务工程重点实验室, 北京 100101
    2. 北京联合大学机器人学院, 北京 100044
  • 收稿日期:2020-03-31 出版日期:2021-08-20 发布日期:2021-08-18
  • 作者简介:梁晔(1978—),女,内蒙古赤峰人,博士,讲师,主要研究方向为图像处理. E-mail: liangye@buu.edu.cn
  • 基金资助:
    国家自然科学基金资助项目(61871038);国家自然科学基金资助项目(61871039);人才强校优选计划领军计划资助项目(BPHR2020AZ02);北京联合大学科研资助项目(ZK30202107)

Image-dependent fusion method for saliency maps

Ye LIANG1,2(),Nan MA2,Hongzhe LIU1   

  1. 1. Beijing Key Laboratory of Information Service Engineering, Beijing Union University, Beijing 100101, China
    2. College of Robotics, Beijing Union University, Beijing 100044, China
  • Received:2020-03-31 Online:2021-08-20 Published:2021-08-18

摘要:

提出基于脊回归的显著图融合方法以获得更好的检测效果。在训练集中寻找待检测图像的近邻图像集,对近邻图像集采用脊回归方法对多种显著性检测方法的融合系数进行估计,进而对不同检测方法的显著图进行融合。该方法充分考虑了检测方法的差异性,很好的解决检测图像在没有基准二值标注下显著图的融合问题。试验采用流行的显著性数据集和显著性检测方法,本研究方法在ECSSD数据集上的AUC为0.911,在HKU-IS数据集上的AUC为0.987, 在DUT-OMRON数据集上的AUC为0.953,结果验证了融合方法的有效性。

关键词: 视觉注意力机制, 显著性检测, 显著性融合, 图像依赖, 脊回归

Abstract:

A saliency fusion method based on ridge regression was proposed to obtain better detection performance. The nearest neighbor set of the image to be detected was searched in the training set. The ridge regression method was used to estimate the fusion coefficients of different saliency maps. The saliency maps of different detection methods were fused. This method fully considered the differences of detection methods, and solved the problem of saliency map fusion in the absence of benchmark binary annotations. The AUC value of the proposed method was 0.911 on ECSSD dataset. The AUC value of the proposed method was 0.987 on HKU-IS dataset. The AUC value of the proposed method was 0.953 on DUT-OMRON dataset. The efficiency of the proposed method was verified by experimental results.

Key words: visual attention mechanism, saliency detection, saliency fusion, image-dependence, ridge regression

中图分类号: 

  • TP391.41

图1

不同方法的检测结果举例"

表1

ECSSD数据集上的检测结果"

方法 F-measure AUC MAE
MR 0.637 0.834 0.237
BSCA 0.658 0.845 0.233
MBS 0.634 0.826 0.231
MC 0.628 0.837 0.251
SELD 0.769 0.866 0.156
DCL 0.699 0.889 0.162
DS 0.717 0.894 0.189
MDF 0.729 0.859 0.174
指数融合方法 0.718 0.901 0.199
线性融合方法 0.598 0.734 0.313
全局融合方法 0.659 0.875 0.332
本研究方法 0.777 0.911 0.171

表2

HKU-IS数据集上的检测结果"

方法 F-measure AUC MAE
MR 0.645 0.867 0.188
BSCA 0.661 0.910 0.175
MBS 0.681 0.920 0.174
MC 0.651 0.909 0.184
SELD 0.799 0.958 0.074
DCL 0.748 0.977 0.072
DS 0.811 0.981 0.079
MDF 0.770 0.969 0.129
指数融合方法 0.773 0.980 0.126
线性融合方法 0.531 0.808 0.309
全局融合方法 0.758 0.915 0.183
本文方法 0.818 0.987 0.071

表3

DUT-OMRON数据集上的检测结果"

方法 F-measure AUC MAE
MR 0.406 0.727 0.246
BSCA 0.574 0.865 0.192
MBS 0.576 0.868 0.157
MC 0.403 0.771 0.238
SELD 0.677 0.914 0.092
DCL 0.629 0.922 0.097
DS 0.675 0.940 0.120
MDF 0.663 0.898 0.093
指数融合方法 0.677 0.943 0.129
线性融合方法 0.448 0.759 0.370
全局融合方法 0.567 0.863 0.239
本文方法 0.688 0.953 0.092

表4

ECSSD数据集上的检测结果"

方法 Smeasure wFmeasure Emeasure
MR 0.639 0.476 0.599
BSCA 0.668 0.495 0.627
MBS 0.632 0.499 0.648
MC 0.640 0.441 0.579
SELD 0.760 0.717 0.805
DCL 0.784 0.698 0.769
DS 0.746 0.618 0.687
MDF 0.712 0.653 0.735
本文方法 0.786 0.721 0.787

表5

HKU-IS数据集上的检测结果"

方法 Smeasure wFmeasure Emeasure
MR 0.669 0.444 0.622
BSCA 0.702 0.467 0.658
MBS 0.714 0.493 0.697
MC 0.687 0.423 0.612
SELD 0.820 0.742 0.876
DCL 0.860 0.739 0.840
DS 0.852 0.709 0.846
MDF 0.810 0.564 0.739
本文方法 0.863 0.736 0.880

表6

DUT-OMRON数据集上的检测结果"

方法 Smeasure wFmeasure Emeasure
MR 0.645 0.380 0.626
BSCA 0.652 0.370 0.629
MBS 0.651 0.419 0.678
MC 0.649 0.347 0.610
SELD 0.750 0.592 0.789
DCL 0.764 0.575 0.750
DS 0.750 0.487 0.704
MDF 0.721 0.565 0.758
本文方法 0.758 0.597 0.793

图2

ECCSD上的结果"

图3

HKU-IS数据集上的结果"

图4

DUT-OMRON数据集上的结果"

1 KOCH Christof , ULLMAN Shimon . Shifts in selective visual attention: towards the underlying neural circuitry[J]. Human Neurbiology, 1985, 4 (4): 219- 227.
2 ITTI Laurent , KOCH Christof , Niebur Ernst . A model of saliency-based visual attention for rapid scene analysis[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1998, 20 (11): 1254- 1259.
doi: 10.1109/34.730558
3 CHENG Mingming , ZHANG Guoxin , MITRA Niloy J , et al. Global contrast based salient region detection[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37 (3): 569- 582.
doi: 10.1109/TPAMI.2014.2345401
4 SHEN Xiaohui, WU Ying. A unified approach to salient object detection via low rank matrix Recovery[C]//Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition. Providence, America: IEEE Press, 2012: 853-860.
5 CHENG Mingming, WARRELL Jonathan, LIN Wenyan, et al. Efficient salient region detection with soft image abstraction[C]//Proceedings of IEEE International Conference on Computer Vision. Sydney, Australia: IEEE Press, 2013: 1529-1536.
6 YAN Qiong, XU Li, SHI Jianping, et al. Hierarchical saliency detection[C]//Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition. Portland, America: IEEE Press, 2013: 1155-1162.
7 MARGOLIN Ran, TAL Ayellet, ZELNIKMANOR Lihi. What makes a patch distinct[C]// Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition. Portland, America: IEEE Press, 2013: 1139-1146.
8 JUDD Tilke M, EHINGER Krista, DURAND Frédo, et al. Learning to predict where humans look[C]//Proceedings of IEEE International Conference on Computer Vision. Kyoto, Japan: IEEE Press, 2009: 2106-2113.
9 BORJI Ali. Boosting bottom-up and top-down visual features for saliency estimation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Rhode Island, America: IEEE Press, 2012, 157(10): 438-445.
10 CHENG Mingming, LIU Yun, LIN Wenyan, et al. Bing: binarized normed gradients for objectness estimation at 300fps[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Columbus, America: IEEE Press, 2014: 3286-3293.
11 WANG Jingdong , JIANG Huaizu , YUAN Zejian , et al. Salient object detection: A discriminative regional feature integration approach[J]. International Journal of Computer Vision, 2017, 123 (2): 251- 268.
doi: 10.1007/s11263-016-0977-3
12 WEI Yichen, WEN Fang, ZHU Wangjiang, et al. Geodesic saliency using background priors[C]//Proceedings of the European Conference on Computer Vision. Firenze, Italy: Springer, 2012: 29-42.
13 MAI Long, NIU Yuzhen, LIU Feng. Saliency aggregation: a data driven approach[C]//Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition. Portland, America: IEEE Press, 2013: 1131-1138.
14 LU Huchuan, RUAN Xiang, YANG Minghsuan. Deep networks for saliency detection via local estimation and global search[C]//Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition. Boston, America: IEEE Press, 2015: 3183-3192.
15 LI Guanbin, YU Yizhou. Visual saliency based on multiscale deep features[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Boston, America: IEEE Press, 2015: 5455-5463.
16 ZHAO Rui, OUYANG Wanli, LI Hongsheng, et al. Saliency detection by multi-context deep learning[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston, America: IEEE Press, 2015: 1265-1274.
17 ZHANG Pingping, WANG Dong, LU Huchuan, et al. Amulet: aggregating multi-level convolutional features for salient object detection[C]//Proceedings of IEEE International Conference on Computer Vision. Venice, Italy: IEEE Press, 2017: 202-211.
18 HOU Qibin, CHENG Mingming, HU Xiaowei, et al. Deeply supervised salient object detection with short connection[C]// Proceedings of IEEE International Conference on Computer Vision and Pattern Recognition. Hawai, America: IEEE Press, 2017: 3203-3212.
19 HU Xiaowei, ZHU Lei, QIN Jing, et al. Recurrently aggregating deep features for salient object detection[C]//Proceedings of Thirty-Second AAAI Conference on Artificial Intelligence. Louisiana, America: AAAI Press, 2018: 6943-6950.
20 LIU Jiangjiang, HOU Qibin, CHENG Mingming, et al. A simple pooling-based design for real-time salient object detection[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. California, America: IEEE Press, 2019: 3917-3926.
21 QIN Xuebin, ZHANG Zichen, HUANG Chenyang, et al. BASNet: boundary-aware salient object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Los Angeles, America: IEEE Press, 2019: 7479-7489.
22 MAI Long, LIU Feng. Comparing salient object detection results without ground truth[C]//Proceedings of European Conference on Computer Vision. Zurich, Switzerland: Springer, 2014: 76-91.
23 LI Guanbin, YU Yizhou. Deep contrast learning for salient object detection[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, America: IEEE Press, 2016: 478-487.
24 LEE Gayoung, TAI Yuwing, KIM Junmo. Deep saliency with encoded low level distance map and high level features[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, America: IEEE Press, 2016: 660-668.
25 LI Xi , ZHAO Liming , WEI Lina , et al. Deep saliency: multi-task deep neural network model for salient object detection[J]. IEEE Transactions on Image Processing, 2016, 25 (8): 3919- 3930.
doi: 10.1109/TIP.2016.2579306
26 JIANG Bowen, ZHANG Lihe, LU Huchuan, et al. Saliency detection via absorbing markov chain[C]//Proceedings of IEEE International Conference on Computer Vision. Sydney, Australia: IEEE Press, 2013: 1665-1672.
27 YANG Chuan, ZHANG Lihe, LU Huchuan, et al. Saliency detection via graph-based manifold ranking[C]//2013 IEEE Conference on Computer Vision and Pattern Recognition. Portlan, America: IEEE Press, 2013: 3166-3173.
28 QIN Yao, LU Huchuan, XU Yiqun. Saliency detection via cellular automata[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Boston, America: IEEE Press, 2015: 110-119.
29 ZHANG Jianming, SCLAROFF Stan, LIN Zhe, et al. Minimum barrier salient object detection at 80 FPS[C]//Proceedings of IEEE International Conference on Computer Vision. Santiago, Chile: IEEE Press, 2015: 1404-1412.
30 MARGOLIN Ran, ZELNIKMANOR Lihi, TAL Ayellet. How to evaluate foreground maps?[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Columbu, America: IEEE Press, 2014: 248-255.
31 FAN Dengping, CHENG Mingming, LIU Yun. Structure- measure: a new way to evaluate foreground maps[C]//Proceedings of the IEEE International Conference on Computer Vision. Venice, Italy: IEEE Press, 2017: 4548-4557.
32 FAN Dengping, GONG Cheng, CAO Yang, et al. Enhanced-alignment measure for binary foreground map evaluation[C]//Proceedings of the International Joint Conference on Artificial Intelligence. Stockholm, Sweden: IJCAI Press, 2018: 698-704.
[1] 任永峰,董学育. 基于自适应流形相似性的图像显著性区域提取算法[J]. 山东大学学报(工学版), 2017, 47(3): 56-62.
[2] 翟继友,周静波,任永峰,王志坚. 基于背景和前景交互传播的图像显著性检测[J]. 山东大学学报(工学版), 2017, 47(2): 80-85.
[3] 任永峰, 周静波. 基于信息弥散机制的图像显著性区域提取算法[J]. 山东大学学报(工学版), 2015, 45(6): 1-6.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 王素玉,艾兴,赵军,李作丽,刘增文 . 高速立铣3Cr2Mo模具钢切削力建模及预测[J]. 山东大学学报(工学版), 2006, 36(1): 1 -5 .
[2] 张永花,王安玲,刘福平 . 低频非均匀电磁波在导电界面的反射相角[J]. 山东大学学报(工学版), 2006, 36(2): 22 -25 .
[3] 李 侃 . 嵌入式相贯线焊接控制系统开发与实现[J]. 山东大学学报(工学版), 2008, 38(4): 37 -41 .
[4] 施来顺,万忠义 . 新型甜菜碱型沥青乳化剂的合成与性能测试[J]. 山东大学学报(工学版), 2008, 38(4): 112 -115 .
[5] 孔祥臻,刘延俊,王勇,赵秀华 . 气动比例阀的死区补偿与仿真[J]. 山东大学学报(工学版), 2006, 36(1): 99 -102 .
[6] 余嘉元1 , 田金亭1 , 朱强忠2 . 计算智能在心理学中的应用[J]. 山东大学学报(工学版), 2009, 39(1): 1 -5 .
[7] 陈瑞,李红伟,田靖. 磁极数对径向磁轴承承载力的影响[J]. 山东大学学报(工学版), 2018, 48(2): 81 -85 .
[8] 王波,王宁生 . 机电装配体拆卸序列的自动生成及组合优化[J]. 山东大学学报(工学版), 2006, 36(2): 52 -57 .
[9] 李可,刘常春,李同磊 . 一种改进的最大互信息医学图像配准算法[J]. 山东大学学报(工学版), 2006, 36(2): 107 -110 .
[10] 季涛,高旭,孙同景,薛永端,徐丙垠 . 铁路10 kV自闭/贯通线路故障行波特征分析[J]. 山东大学学报(工学版), 2006, 36(2): 111 -116 .