高级搜索

基于样本选择的RGBD图像协同显著目标检测

刘政怡 刘俊雷 赵鹏

引用本文: 刘政怡, 刘俊雷, 赵鹏. 基于样本选择的RGBD图像协同显著目标检测[J]. 电子与信息学报, doi: 10.11999/JEIT190393 shu
Citation:  Zhengyi LIU, Junlei LIU, Peng ZHAO. RGBD Image Co-saliency Object Detection Based on Sample Selection[J]. Journal of Electronics and Information Technology, doi: 10.11999/JEIT190393 shu

基于样本选择的RGBD图像协同显著目标检测

    作者简介: 刘政怡: 女,1978年生,副教授,研究方向为计算机视觉、深度学习;
    刘俊雷: 男,1995年生,硕士生,研究方向为计算机视觉;
    赵鹏: 女,1976年生,副教授,研究方向为智能信息处理、机器学习
    通讯作者: 刘政怡,liuzywen@ahu.edu.cn
  • 基金项目: 安徽省自然科学基金(1908085MF182),国家自然科学基金(61602004)

摘要: 协同显著目标检测的目的是在包含两张及以上相关图像的图像组中检测共同显著的物体。该文提出一种利用机器学习的方法对协同显著目标进行检测。首先,基于4个评分指标从图像组中选择部分显著目标易于检测的简单图像,构成简单图像集;接着,基于协同一致性的原则,从简单图像集中提取正负样本,并用深度学习模型提取的高维语义特征表示正负样本;再者,利用正负样本训练的协同显著分类器对图像中的超像素进行分类,得到协同显著目标区域;最后,经过一个平滑融合的操作,得到最终的协同显著图。在公开数据集上的测试结果表明,所提算法在检测精度和检测效率上优于目前的主流算法,并具有较强的鲁棒性。

English

    1. [1]

      WANG Wenguan, SHEN Jianbing, LI Xuelong, et al. Robust video object cosegmentation[J]. IEEE Transactions on Image Processing, 2015, 24(10): 3137–3148. doi: 10.1109/TIP.2015.2438550

    2. [2]

      LEI Jianjun, WU Min, ZHANG Changqing, et al. Depth-preserving stereo image retargeting based on pixel fusion[J]. IEEE Transactions on Multimedia, 2017, 19(7): 1442–1453. doi: 10.1109/TMM.2017.2660440

    3. [3]

      LI Chongyi, GUO Jichang, CONG Runmin, et al. Underwater image enhancement by dehazing with minimum information loss and histogram distribution prior[J]. IEEE Transactions on Image Processing, 2016, 25(12): 5664–5677. doi: 10.1109/TIP.2016.2612882

    4. [4]

      CAO Xiaochun, ZHANG Changqing, FU Huazhu, et al. Saliency-aware nonparametric foreground annotation based on weakly labeled data[J]. IEEE Transactions on Neural Networks and Learning Systems, 2016, 27(6): 1253–1265. doi: 10.1109/TNNLS.2015.2488637

    5. [5]

      PANG Yanwei, ZHU Hailong, LI Xuelong, et al. Motion blur detection with an indicator function for surveillance machines[J]. IEEE Transactions on Industrial Electronics, 2016, 63(9): 5592–5601. doi: 10.1109/TIE.2016.2564938

    6. [6]

      LEI Jianjun, LIU Jianying, ZHANG Hailong, et al. Motion and structure information based adaptive weighted depth video estimation[J]. IEEE Transactions on Broadcasting, 2015, 61(3): 416–424. doi: 10.1109/TBC.2015.2437197

    7. [7]

      YANG Jingyu, GAN Ziqiao, LI Kun, et al. Graph-based segmentation for RGB-D data using 3-D geometry enhanced superpixels[J]. IEEE Transactions on Cybernetics, 2015, 45(5): 927–940. doi: 10.1109/TCYB.2014.2340032

    8. [8]

      SONG Hangke, LIU Zhi, XIE Yufeng, et al. RGBD co-saliency detection via bagging-based clustering[J]. IEEE Signal Processing Letters, 2016, 23(12): 1722–1726. doi: 10.1109/LSP.2016.2615293

    9. [9]

      CONG Runmin, LEI Jianjun, FU Huazhu, et al. Co-saliency detection for RGBD images based on multi-constraint feature matching and cross label propagation[J]. IEEE Transactions on Image Processing, 2018, 27(2): 568–579. doi: 10.1109/TIP.2017.2763819

    10. [10]

      CONG Runmin, LEI Jianjun, FU Huazhu, et al. An iterative co-saliency framework for RGBD images[J]. IEEE Transactions on Cybernetics, 2019, 49(1): 233–246. doi: 10.1109/tcyb.2017.2771488

    11. [11]

      CHEN M, VELASCO-FORERO S, TSANG I, et al. Objects co-segmentation: Propagated from simpler images[C]. 2015 IEEE International Conference on Acoustics, Speech and Signal Processing, Brisbane, Australia, 2015: 1682–1686. doi: 10.1109/ICASSP.2015.7178257.

    12. [12]

      ACHANTA R, SHAJI A, SMITH K, et al. SLIC superpixels compared to state-of-the-art superpixel methods[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2012, 34(11): 2274–2282. doi: 10.1109/TPAMI.2012.120

    13. [13]

      GUO Jingfan, REN Tongwei, and BEI Jia. Salient object detection for RGB-D image via saliency evolution[C]. 2016 IEEE International Conference on Multimedia and Expo, Seattle, USA, 2016: 1–6.

    14. [14]

      CONG Runmin, LEI Jianjun, ZHANG Changqing, et al. Saliency detection for stereoscopic images based on depth confidence analysis and multiple cues fusion[J]. IEEE Signal Processing Letters, 2016, 23(6): 819–823. doi: 10.1109/lsp.2016.2557347

    15. [15]

      MAI Long and LIU Feng. Comparing salient object detection results without ground truth[C]. The 13th European Conference on Computer Vision, Zurich, Switzerland, 2014: 76–91. doi: 10.1007/978-3-319-10578-9_6.

    16. [16]

      LI Lina, LIU Zhi, and ZHANG Jian. Unsupervised image co-segmentation via guidance of simple images[J]. Neurocomputing, 2018, 275: 1650–1661. doi: 10.1016/j.neucom.2017.10.002

    17. [17]

      CONG Runmin, LEI Jianjun, FU Huazhu, et al. HSCS: Hierarchical Sparsity based co-saliency detection for RGBD images[J]. IEEE Transactions on Multimedia, 2019, 21(7): 1660–1771. doi: 10.1109/TMM.2018.2884481

    18. [18]

      ARTHUR D and VASSILVITSKII S. k-means++: The advantages of careful seeding[C]. The Annual ACM-SIAM Symposium on Discrete Algorithms, New Orleans, USA, 2007: 1027–1035.

    19. [19]

      HUANG Posheng, SHEN C H, and HSIAO H F. RGBD salient object detection using spatially coherent deep learning framework[C]. The 23rd IEEE International Conference on Digital Signal Processing, Shanghai, China, 2018: 1–5.

    20. [20]

      LIU Zhengyi, SHI Song, DUAN Quntao, et al. Salient object detection for RGB-D image by single stream recurrent convolution neural network[J]. Neurocomputing, 2019, 363: 46–57. doi: 10.1016/j.neucom.2019.07.012

    21. [21]

      HAN Junwei, CHENG Gong, LI Zhenpeng, et al. A unified metric learning-based framework for co-saliency detection[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2018, 28(10): 2473–2483. doi: 10.1109/tcsvt.2017.2706264

    22. [22]

      QIN Yao, FENG Mengyang, LU Huchuan, et al. Hierarchical cellular automata for visual saliency[J]. International Journal of Computer Vision, 2018, 126(7): 751–770. doi: 10.1007/s11263-017-1062-2

    23. [23]

      OTSU N. A threshold selection method from gray-level histograms[J]. IEEE Transactions on Systems, Man, and Cybernetics, 1979, 9(1): 62–66. doi: 10.1109/TSMC.1979.4310076

    24. [24]

      BORJI A, CHENG Mingming, JIANG Huaizu, et al. Salient object detection: A benchmark[J]. IEEE Transactions on Image Processing, 2015, 24(12): 5706–5722. doi: 10.1109/TIP.2015.2487833

    25. [25]

      WANG Wenguan, SHEN Jianbing, and SHAO Ling. Consistent video saliency using local gradient flow optimization and global refinement[J]. IEEE Transactions on Image Processing, 2015, 24(11): 4185–4196. doi: 10.1109/TIP.2015.2460013

    26. [26]

      FAN Dengping, CHENG Mingming, LIU Yun, et al. Structure-measure: A new way to evaluate foreground maps[C]. 2017 IEEE International Conference on Computer Vision, Venice, Italy, 2017: 4558–4567.

    27. [27]

      LI Yijun, FU Keren, LIU Zhi, et al. Efficient saliency-model-guided visual co-saliency detection[J]. IEEE Signal Processing Letters, 2015, 22(5): 588–592. doi: 10.1109/LSP.2014.2364896

    28. [28]

      FU Huazhu, CAO Xiaochun, and TU Zhuowen. Cluster-based co-saliency detection[J]. IEEE Transactions on Image Processing, 2013, 22(10): 3766–3778. doi: 10.1109/TIP.2013.2260166

    1. [1]

      张文明, 姚振飞, 高雅昆, 李海滨. 一种平衡准确性以及高效性的显著性目标检测深度卷积网络模型. 电子与信息学报,

    2. [2]

      付晓薇, 杨雪飞, 陈芳, 李曦. 一种基于深度学习的自适应医学超声图像去斑方法. 电子与信息学报,

    3. [3]

      张惊雷, 厚雅伟. 基于改进循环生成式对抗网络的图像风格迁移. 电子与信息学报,

    4. [4]

      柳长源, 王琪, 毕晓君. 基于多通道多尺度卷积神经网络的单幅图像去雨方法. 电子与信息学报,

    5. [5]

      游凌, 李伟浩, 张文林, 王科人. 基于深度神经网络的Morse码自动译码算法. 电子与信息学报,

    6. [6]

      陈家祯, 吴为民, 郑子华, 叶锋, 连桂仁, 许力. 基于虚拟光学的视觉显著目标可控放大重建. 电子与信息学报,

    7. [7]

      胡永健, 高逸飞, 刘琲贝, 廖广军. 基于图像分割网络的深度假脸视频篡改检测. 电子与信息学报,

    8. [8]

      黄静琪, 胡琛, 孙山鹏, 高翔, 何兵. 一种基于异步传感器网络的空间目标分布式跟踪方法. 电子与信息学报,

    9. [9]

      王一宾, 裴根生, 程玉胜. 基于标记密度分类间隔面的组类属属性学习. 电子与信息学报,

    10. [10]

      唐伦, 曹睿, 廖皓, 王兆堃. 基于深度强化学习的服务功能链可靠部署算法. 电子与信息学报,

    11. [11]

      陈卓, 冯钢, 何颖, 周杨. 运营商网络中基于深度强化学习的服务功能链迁移机制. 电子与信息学报,

    12. [12]

      申铉京, 沈哲, 黄永平, 王玉. 基于非局部操作的深度卷积神经网络车位占用检测算法. 电子与信息学报,

    13. [13]

      赵斌, 王春平, 付强. 显著性背景感知的多尺度红外行人检测方法. 电子与信息学报,

    14. [14]

      陈前斌, 管令进, 李子煜, 王兆堃, 杨恒, 唐伦. 基于深度强化学习的异构云无线接入网自适应无线资源分配算法. 电子与信息学报,

    15. [15]

      周牧, 李垚鲆, 谢良波, 蒲巧林, 田增山. 基于多核最大均值差异迁移学习的WLAN室内入侵检测方法. 电子与信息学报,

    16. [16]

      张凯, 陈彬, 许志伟. 基于多目标进化策略算法的DNA核酸编码设计. 电子与信息学报,

    17. [17]

      熊伟, 顾祥岐, 徐从安, 崔亚奇. 多编队目标先后出现时的无先验信息跟踪方法. 电子与信息学报,

    18. [18]

      项厚宏, 陈伯孝, 杨婷, 杨明磊. 基于多帧相位增强的米波雷达低仰角目标DOA估计方法. 电子与信息学报,

    19. [19]

      王彬, 侯越圣. 一种Alpha稳定分布噪声下的体积目标长度估计方法. 电子与信息学报,

    20. [20]

      姜文, 牛杰, 吴一戎, 梁兴东. 机载多通道SAR运动目标方位向速度和法向速度联合估计算法. 电子与信息学报,

  • 图 1  本文提出的RGBD协同显著目标检测方法的框架图

    图 2  不同方法生成的协同显著图对比

    图 3  本文算法与其他算法在两个数据集上的P-R曲线对比

    图 4  本文算法两个策略在两个数据集上的P-R曲线对比

    图 5  RGBD CoSal150数据集不同参数的F-measure测量

    表 1  不同算法在两个数据集上的测试结果对比

    RGBD CoSal150RGBD CoSeg183
    S-measureF-measureMAES-measureF-measureMAE
    ESCS0.6250.5870.2180.6360.4140.156
    CBCS0.5720.5820.2150.6220.3650.116
    ICFS0.7100.7640.1790.6300.4430.163
    MCL0.7660.8100.1370.6890.4880.098
    本文方法0.8490.8810.0890.7080.5020.081
    下载: 导出CSV

    表 2  不同模块在两个数据集上的测试结果对比

    RGBD CoSal150RGBD CoSeg183
    S-measureF-measureMAES-measureF-measureMAE
    颜色+纹理特征0.8160.8170.1310.6610.4730.143
    无简单图像选择0.8320.8370.1170.7020.4770.090
    高维语义特征+简单图像选择0.8490.8810.0890.7080.5020.081
    下载: 导出CSV

    表 3  不同方法平均每副图运行时间比较(s)

    方法ESCSCBCSICFSMCL本文方法
    时间2.842.4342.6741.038.76
    下载: 导出CSV
  • 加载中
图(5)表(3)
计量
  • PDF下载量:  13
  • 文章访问数:  174
  • HTML全文浏览量:  113
文章相关
  • 通讯作者:  刘政怡, liuzywen@ahu.edu.cn
  • 收稿日期:  2019-06-03
  • 录用日期:  2020-03-01
  • 网络出版日期:  2020-06-27
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

/

返回文章