结合稀疏重构与能量方程优化的显著性检测
|
摘要
针对现有算法在复杂背景图像显著目标检测中存在背景被错误凸显的问题,为抑制背景提取更加准确的前景,提出一种结合稀疏重构与能量优化的显著性检测算法。首先将输入图像分割为超像素以去除不必要的细节;然后选取图像边界超像素作为背景模板,利用其作为稀疏字典计算重构误差,并作为超像素初始显著值;最后构造新的能量方程对初始显著值优化,并在优化后对其前景增强生成最终显著图。在包含真值图像的MSRA10K和ECSSD1000数据集上,将提出的算法与其他10种算法进行对比测试,PR曲线图、准确率P、F值的效果优于其他10种算法的结果。实验结果表明,所提算法在复杂背景图像的显著目标检测中,相比于已有的多种算法鲁棒性更好,能够对背景进行有效的抑制,提取显著目标也更加精确。
As extant method wrongly highlights backgrounds in salient object detection from complex background images,this paper proposed a new algorithm of saliency detection to suppress background in combination with sparse reconstruction and energy optimization,which extracted foreground more accurate instantaneously. Firstly,it decomposed the input image into super pixel within abstracting unnecessary detail. Then,it selected image boundary super pixels as background templates,which used as sparse dictionaries to calculate reconstruction errors that as super pixels initial saliency. Finally,it introduced the energy equation to optimize the initial saliency,and generated the final saliency map after enhanced the foreground of optimized saliency. It tested the proposed algorithm and other 10 algorithms on MSRA10 K and ECSSD1000 dataset with ground truths.The PR curve,precision( P) and F-measure( F) of the proposed algorithm had better performance than other 10 algorithms.The experimental results show that the proposed algorithm is more robust to suppress background effectively,and the extraction of salient object is more precise.
As extant method wrongly highlights backgrounds in salient object detection from complex background images,this paper proposed a new algorithm of saliency detection to suppress background in combination with sparse reconstruction and energy optimization,which extracted foreground more accurate instantaneously. Firstly,it decomposed the input image into super pixel within abstracting unnecessary detail. Then,it selected image boundary super pixels as background templates,which used as sparse dictionaries to calculate reconstruction errors that as super pixels initial saliency. Finally,it introduced the energy equation to optimize the initial saliency,and generated the final saliency map after enhanced the foreground of optimized saliency. It tested the proposed algorithm and other 10 algorithms on MSRA10 K and ECSSD1000 dataset with ground truths.The PR curve,precision( P) and F-measure( F) of the proposed algorithm had better performance than other 10 algorithms.The experimental results show that the proposed algorithm is more robust to suppress background effectively,and the extraction of salient object is more precise.
引文
[1] Zheng Liang,Wang Shengjin,Liu Ziqion,et al. Fast image retrieval:query pruning and early termination[J]. IEEE Trans on Multimedia,2015,17(5):648-659.
[2] Liu Zhi,Shi Ran,Shen Liquan,et al. Unsupervised salient object segmentation based on kernel density estimation and two-phase graph cut[J]. IEEE Trans on Multimedia,2012,14(4):1275-1289.
[3] Shen Liquan,Liu Zhi,Zhang Zhaoyang. A novel H. 264 rate control algorithm with consideration of visual attention[J]. Multimedia Tools&Applications,2013,63(3):709-727.
[4] Chen Zhihui,Wang Hanzi,Zhang Liming,et al. Visual saliency detection based on homology similarity and an experimental evaluation[J]. Journal of Visual Communication&Image Representation,2016,40(10):251-264.
[5] Lie M I,Borba G B,Neto H V,et al. Fast saliency detection using sparse random color samples and joint upsampling[C]//Proc of the29th SIBGRAPI Conference on Graphics, Patterns and Images.Washington DC:IEEE Computer Society,2017:217-224.
[6]郑重,姚婷婷,孙永宣,等.基于高层颜色语义名称的显著性检测[J].计算机应用研究,2018,35(7):2197-2199,2232.(Zheng Zhong,Yao Tingting,Sun Yongxuan,et al. Saliency detection based on high-level color semantic names[J]. Application Research of Computers,2018,35(7):2197-2199,2232.)
[7] Itti L,Koch C,Niebur E. A model of saliency-based visual attention for rapid scene analysis[J]. IEEE Trans on Pattern Analysis&Machine Intelligence,1998,20(11):1254-1259.
[8] Li Xi,Li Yao,Shen Chunhua,et al. Contextual hypergraph modeling for salient object detection[C]//Proc of IEEE International Conference on Computer Vision. Washington DC:IEEE Computer Society,2014:3328-3335.
[9] Xie Yulin,Lu Huchuan,Yang M H. Bayesian saliency via low and mid level cues[J]. IEEE Trans on Image Processing,2013,22(5):1689-1698.
[10]Achanta R,Hemami S,Estrada F,et al. Frequency-tuned salient region detection[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2009:1597-1604.
[11]Margolin R,Tal A,Zelnik-Manor L. What makes a patch distinct?[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2013:1139-1146.
[12]Hornung A,Pritch Y,Krahenbuhl P,et al. Saliency filters:contrast based filtering for salient region detection[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. New York:ACM Press,2012:733-740.
[13]Cheng Mingming,Mitra N J,Huang Xiaolei,et al. Global contrast based salient region detection[J]. IEEE Trans on Pattern Analysis&Machine Intelligence,2015,37(3):569-582.
[14]Li Xiaohui,Lu Huchuan,Zhang Lihe,et al. Saliency detection via dense and sparse reconstruction[C]//Proc of IEEE International Conference on Computer Vision. Washington DC:IEEE Computer Society,2013:2976-2983.
[15]Yang Chuan,Zhang Lihe,Lu Huchuan,et al. Saliency detection via graph-based manifold ranking[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Washington DC:IEEE Computer Society,2013:3166-3173.
[16]Zhu Wangjiang,Liang Shuang,Wei Yichen,et al. Saliency optimization from robust background detection[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Washington DC:IEEE Computer Society,2014:2814-2821.
[17]Achanta R,Shaji A,Smith K,et al. SLIC superpixels compared to state-of-the-art superpixel methods[J]. IEEE Trans on Pattern Analysis&Machine Intelligence,2012,34(11):2274-2282.
[18]Zhang Geng,Yuan Zejian,Liu Yuehu,et al. Video object segmentation by integrating trajectories from points and regions[J]. Multimedia Tools&Applications,2015,74(21):9665-9696.
[19]崔丽群,吴晓冬,赵越.一种基于CRF-MR的自顶向下显著性目标检测方法[J].计算机应用研究,2018,35(8):2535-2539.(Cui Liqun,Wu Xiaodong,Zhao Yue. Top-down saliency target detection based on CRF-MR[J]. Application Research of Computers,2018,35(8):2535-2539.)
[20]张巧荣.利用背景先验的显著性检测算法[J].中国图象图形学报,2016,21(2):165-173.(Zhang Qiaorong. Saliency detection algorithm based on background prior[J]. Journal of Image&Graphics,2016,21(2):165-173.)
[21]Borji A,Cheng Mingming,Jiang Huaizu,et al. Salient object detection:a benchmark[J]. IEEE Trans on Image Processing,2015,24(12):5706-5722.
[2] Liu Zhi,Shi Ran,Shen Liquan,et al. Unsupervised salient object segmentation based on kernel density estimation and two-phase graph cut[J]. IEEE Trans on Multimedia,2012,14(4):1275-1289.
[3] Shen Liquan,Liu Zhi,Zhang Zhaoyang. A novel H. 264 rate control algorithm with consideration of visual attention[J]. Multimedia Tools&Applications,2013,63(3):709-727.
[4] Chen Zhihui,Wang Hanzi,Zhang Liming,et al. Visual saliency detection based on homology similarity and an experimental evaluation[J]. Journal of Visual Communication&Image Representation,2016,40(10):251-264.
[5] Lie M I,Borba G B,Neto H V,et al. Fast saliency detection using sparse random color samples and joint upsampling[C]//Proc of the29th SIBGRAPI Conference on Graphics, Patterns and Images.Washington DC:IEEE Computer Society,2017:217-224.
[6]郑重,姚婷婷,孙永宣,等.基于高层颜色语义名称的显著性检测[J].计算机应用研究,2018,35(7):2197-2199,2232.(Zheng Zhong,Yao Tingting,Sun Yongxuan,et al. Saliency detection based on high-level color semantic names[J]. Application Research of Computers,2018,35(7):2197-2199,2232.)
[7] Itti L,Koch C,Niebur E. A model of saliency-based visual attention for rapid scene analysis[J]. IEEE Trans on Pattern Analysis&Machine Intelligence,1998,20(11):1254-1259.
[8] Li Xi,Li Yao,Shen Chunhua,et al. Contextual hypergraph modeling for salient object detection[C]//Proc of IEEE International Conference on Computer Vision. Washington DC:IEEE Computer Society,2014:3328-3335.
[9] Xie Yulin,Lu Huchuan,Yang M H. Bayesian saliency via low and mid level cues[J]. IEEE Trans on Image Processing,2013,22(5):1689-1698.
[10]Achanta R,Hemami S,Estrada F,et al. Frequency-tuned salient region detection[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2009:1597-1604.
[11]Margolin R,Tal A,Zelnik-Manor L. What makes a patch distinct?[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Piscataway,NJ:IEEE Press,2013:1139-1146.
[12]Hornung A,Pritch Y,Krahenbuhl P,et al. Saliency filters:contrast based filtering for salient region detection[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. New York:ACM Press,2012:733-740.
[13]Cheng Mingming,Mitra N J,Huang Xiaolei,et al. Global contrast based salient region detection[J]. IEEE Trans on Pattern Analysis&Machine Intelligence,2015,37(3):569-582.
[14]Li Xiaohui,Lu Huchuan,Zhang Lihe,et al. Saliency detection via dense and sparse reconstruction[C]//Proc of IEEE International Conference on Computer Vision. Washington DC:IEEE Computer Society,2013:2976-2983.
[15]Yang Chuan,Zhang Lihe,Lu Huchuan,et al. Saliency detection via graph-based manifold ranking[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Washington DC:IEEE Computer Society,2013:3166-3173.
[16]Zhu Wangjiang,Liang Shuang,Wei Yichen,et al. Saliency optimization from robust background detection[C]//Proc of IEEE Conference on Computer Vision and Pattern Recognition. Washington DC:IEEE Computer Society,2014:2814-2821.
[17]Achanta R,Shaji A,Smith K,et al. SLIC superpixels compared to state-of-the-art superpixel methods[J]. IEEE Trans on Pattern Analysis&Machine Intelligence,2012,34(11):2274-2282.
[18]Zhang Geng,Yuan Zejian,Liu Yuehu,et al. Video object segmentation by integrating trajectories from points and regions[J]. Multimedia Tools&Applications,2015,74(21):9665-9696.
[19]崔丽群,吴晓冬,赵越.一种基于CRF-MR的自顶向下显著性目标检测方法[J].计算机应用研究,2018,35(8):2535-2539.(Cui Liqun,Wu Xiaodong,Zhao Yue. Top-down saliency target detection based on CRF-MR[J]. Application Research of Computers,2018,35(8):2535-2539.)
[20]张巧荣.利用背景先验的显著性检测算法[J].中国图象图形学报,2016,21(2):165-173.(Zhang Qiaorong. Saliency detection algorithm based on background prior[J]. Journal of Image&Graphics,2016,21(2):165-173.)
[21]Borji A,Cheng Mingming,Jiang Huaizu,et al. Salient object detection:a benchmark[J]. IEEE Trans on Image Processing,2015,24(12):5706-5722.