基于线性融合的流行排序显著性检测算法的研究
|
摘要
目的:针对现有的图的流行排序显著性检测算法忽略多尺度超像素之间的空间信息,而造成过分依赖某一种超像素分割的问题,本文提出一种从多角度考虑的显著性检测算法。方法:首先对图像进行多尺度超像素分割,然后利用经典的流行排序算法对分割后的图像分别计算单层显著图,最后利用层次关系融合多层显著图,得到最终显著图。结果:实验结果表明,本文算法获得更高的精确率,优于传统显著性检测算法。
The goal of the saliency detection is to detect important regions in an image,existing graph-based manifold ranking saliency detection method is less effective due to neglecting the spatial information between multiscale superpixels, lead to excessive dependence on one kind of super pixel segmentation, an improved method is proposed to got saliency map from multi layers. First, the image is segmented by multi-scale super pixel. Then,the classical popular sorting algorithm is applied to compute the single layer saliency map of the segmented images. Finally, the final saliency map is obtained by using the hierarchical relation to fuse the multi-layer saliency map. The experimental comparison results demonstrate the effectiveness and superiority of the proposed improved method.
The goal of the saliency detection is to detect important regions in an image,existing graph-based manifold ranking saliency detection method is less effective due to neglecting the spatial information between multiscale superpixels, lead to excessive dependence on one kind of super pixel segmentation, an improved method is proposed to got saliency map from multi layers. First, the image is segmented by multi-scale super pixel. Then,the classical popular sorting algorithm is applied to compute the single layer saliency map of the segmented images. Finally, the final saliency map is obtained by using the hierarchical relation to fuse the multi-layer saliency map. The experimental comparison results demonstrate the effectiveness and superiority of the proposed improved method.
引文
[1]L.Itti,C.Koch,and E.Niebur.A model of saliency-based visual attention for rapid scene analysis[J],IEEE Transactions on Pattern Analysis and Machine Intelligence,1998:1254-1259,20(11).
[2]T.Liu,J.Sun,N.Zheng,X.Tang,and H.-Y.Shum.Learning to detect a salient object[C].Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition,2007:1-8.
[3]Y.-F.Ma and H.-J.Zhang.Contrast-based image attention analysis by using fuzzy growing[J].In ACM international Conference on Multimedia(ACM Multimedia),2003.
[4]C.Yang,L.Zhang,H.Lu,X.Ruan,and M.-H.Yang.Saliency detection via graph-based manifold ranking[C].Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition,2013:3166-3173.
[2]T.Liu,J.Sun,N.Zheng,X.Tang,and H.-Y.Shum.Learning to detect a salient object[C].Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition,2007:1-8.
[3]Y.-F.Ma and H.-J.Zhang.Contrast-based image attention analysis by using fuzzy growing[J].In ACM international Conference on Multimedia(ACM Multimedia),2003.
[4]C.Yang,L.Zhang,H.Lu,X.Ruan,and M.-H.Yang.Saliency detection via graph-based manifold ranking[C].Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition,2013:3166-3173.