摘要
依据图像底层的颜色对比度特征,提出一种自底向上、数据驱动的视觉显著性检测方法,从全局对比度考虑,提取显著性目标的全分辨率显著图。首先通过对图像局部结构特征的分析得到关于目标和背景的先验分布信息;在分布信息的基础上分别提取图像的全局颜色对比度特征和空间位置关系特征,以空间关系权重优化显著性检测结果;进一步融合频域谱残差显著图,降低背景冗余及弱小显著目标对全局显著性检测结果的影响。在国际公开的显著性测试数据集MSRA-1000上进行实验,结果表明:该方法由于抑制了非显著区域的干扰,相对于现有的一些方法更能突出复杂背景下的显著目标。
Based on the low-level features of the image color contrast,we proposed a bottom-up data driven algorithm of visual saliency detection. According to global contrast,our model generated a fullresolution saliency map. First of all,by analyzing the characteristics of the local structure of the image,the distribution prior knowledge of the object and the background was obtained. On the basis of the distribution information,we extracted the global color contrast and spatial position features that could be used to significantly optimize the detection performance. Furthermore,we integrated the Spectral Residual( SR) saliency map on the purpose of reducing the impact of background redundancy and small regions with high saliency. And due to the suppression of background,results of the experiments on public saliency detection database( MSRA-1000) showed that the presented methodcould effectively highlight the salient object region in complex background compared to some existing methods.
引文
[1]Jiang P,Ling H,Yu J,et al.Salient Region Detection by UFO:Uniqueness,Focusness and Objectness[C]//IEEE International Conference on Computer Vision.USA:[s.n.],2013:1976-1983.
[2]Itti L,Koch C,Niebur E.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.
[3]Hou X,Zhang L.Saliency detection:A spectral residual approach[C]//IEEE Conference on Computer Vision and Pattern Recognition.USA:[s.n.],2007:1-8.
[4]Zhai Y,Shah M.Visual attention detection in video sequences using spatiotemporal cues[C]//ACM International Conference on Multimedia.USA:[s.n.],2005:815-824.
[5]Cheng M,Zhang G,Mitra N,et al.Global contrast based salient region detection[C]//IEEE Conference on Computer Vision and Pattern Recognition.USA:[s.n.],2011:409-416.
[6]Achanta R,Hemami S,Estrada F,et al.Frequency-tuned salient region detection[C]//IEEE Conference on Computer Vision and Pattern Recognition.USA:[s.n.],2009:1597-1604.
[7]Perazzi F,Krahenbuhl P,Pritch Y,et al.Saliency filters:Contrast based filtering for salient region detection[C]//IEEE Conference on Computer Vision and Pattern Recognition.USA:[s.n.],2012:733-740.
[8]Jiang Z,Davis L S.Submodular salient region detection[C]//IEEE Conference on Computer Vision and Pattern Recognition.USA:[s.n.],2013:2043-2050.
[9]李崇飞,高颖慧,卢凯,等.基于结构相似度的视觉显著性检测方法[J].计算机工程与科学,2013,35(10):181-185.
[10]Weickert J.Anisotropic Diffusion in Image Processing[M].Germany:Teubner,Stuttgart,1988.
[11]邵宇,孙富春,刘莹.基于局部结构张量的无参考型图像质量评价方法[J].电子与信息学报,2012,34(8):1779-1785.
[12]Junq C,Kim C.A unified spectral-domain approach for saliency detection and its application to automatic object segmentation[J].IEEE Transactions on Image Processing,2012,21(3):1272-1283.
[13]Zhou L,Yang Z.Salient region detection based on spatial and background priors[C]//IEEE International Conference on Information and Automation.USA:[s.n.],2014:262-266.
[14]Harel J,Koch C,Perona P.Graph-based visual saliency[J].Advances in Neural Information Processing Systems,2006,19:545-552.
[15]Achanta R,Estrada F,Wils P,et al.Salient region detection and segmentation[C]//IEEE Conference on Computer Vision Systems.USA:[s.n.],2008:66-75.