基于二维离散分数阶傅里叶变换的视觉注意力算法
|
摘要
为了改进现有几种视觉注意力算法存在抗干扰性差的不足,提出了一种基于分数阶傅里叶变换(Fr FT)的视觉注意力算法.首先从信息熵的角度对图像背景中的噪声进行分析,以获取前景目标为目的,利用傅里叶变换中的分数阶与噪声的密度分布进行加权作为显著性图的最终权值,同时依据熵直方图对所获得的显著性图进行优化.最后利用MSRA10K图像库作为输入数据,分析查准率-查全率与受试者工作特征(ROC)曲线.实验结果表明,文中提出的算法有更好的屏蔽噪声能力与自适应能力,在一定程度上能确切地模拟人类视觉.
In order to improve the lack of the anti-noise performance of existing visual attention algorithms,a visual attention algorithm based on fractional Fourier transform(Fr FT) is proposed. Firstly,the background noise of image is analyzed in accordance with information entropy,which aims at obtaining the foreground target. The weighting is conducted as the final weight value of the saliency map by means of the distribution of fractional order and noise density in fractional Fourier transform. At the same time,the saliency maps are optimized on the basis of the entropy histogram. Finally,the image database MSRA10 K is used as input data to analyze precision-recall and ROC curves.Experiments show that the proposed algorithm can simulate human fixation to some extent.
In order to improve the lack of the anti-noise performance of existing visual attention algorithms,a visual attention algorithm based on fractional Fourier transform(Fr FT) is proposed. Firstly,the background noise of image is analyzed in accordance with information entropy,which aims at obtaining the foreground target. The weighting is conducted as the final weight value of the saliency map by means of the distribution of fractional order and noise density in fractional Fourier transform. At the same time,the saliency maps are optimized on the basis of the entropy histogram. Finally,the image database MSRA10 K is used as input data to analyze precision-recall and ROC curves.Experiments show that the proposed algorithm can simulate human fixation to some extent.
引文
[1] SCHLKOPF B,PLATT J,HOFMANN T. Graph-based visual saliency[J]. Nips,2007,19:545-552.
[2] MEUR O L,CALLET L,BARBA D,et al. A coherent computational approach to model bottom-up visual attention[J]. IEEE Transactions on Pattern Analysis&Ma-chine Intelligence,2006,28(5):802-817.
[3] GUO C,MA Q,ZHANG L. Spatio-temporal saliency detection using phase spectrum of quaternion Fourier transform[C]∥Proceedings of 2008 IEEE Conference on Computer Vision and Pattern Recognition. Anchorage:IEEE,2008:1-8.
[4] ITTI L,KOCH C,NIEBUR E. A model of saliency-based visual attention for rapid scene analysis[J]. IEEE Transactions on Pattern Analysis&Machine Intelligence,1998,20(11):1254-1259.
[5] ZHANG L,TONG M H,MARKS T K,et al. SUN:a Bayesian framework for saliency using natural statistics[J]. Journal of Vision,2008,8(7):32. 1-32. 20.
[6] HOU X D,ZHANG L Q. Saliency detection:a spectral residual approach[C]∥Proceedings of 2007 IEEE Conference on Computer Vision and Pattern Recognition. Minneapolis:IEEE,2007:1-8.
[7] LI J,LEVINE M D,AN X,et al. Visual saliency based on scale-space analysis in the frequency domain[J]. IEEE Transactions on Pattern Analysis&Machine Intelligence,2013,35(4):996-1010.
[8] JIANG B W,ZHANG L H,LU H C,et al. Saliency detection via absorbing Markov chain[C]∥Proceedings of2013 IEEE International Conference on Computer Vision.Sydney:IEEE,2013:1665-1672.
[9] SUN J,LU H,LIU X. Saliency region detection based on Markov absorption probabilities[J]. IEEE Transactions on Image Processing,2015,24(5):1639-1649.
[10] CHEN S,ZHENG L,HU X,et al. Discriminative saliency propagation with sink points[J]. Pattern Recognition,2016,60:2-12.
[11] SINGH A K,SAXENA R. On convolution and product theorems for FRFT[J]. Wireless Personal Communications,2012,65(1):189-201.
[12] WANG P G,TIAN H,ZHENG W. A novel image fusion method based on FRFT-NSCT[J]. Mathematical Problems in Engineering,2013,2013(11):408232/1-9.
[13] GAO L,QI L,WANG Y J,et al. Rotation invariance in2D-FRFT with application to digital image watermarking[J]. Journal of Signal Processing Systems,2013,72(2):133-148.
[14] WU J H,LUO X Z,ZHOU N R. Four-image encryption method based on spectrum truncation,chaos and the MODFrFT[J]. Optics&Laser Technology,2013,45(1):571-577.
[15] WANG X,QI L,TIE Y,et al. Face recognition based on weighted multi-order feature fusion of 2D-FrFT[M]∥Digital TV and Wireless Multimedia Communication:the13th International Forum. Shanghai:Springer,2016:137-145.
[2] MEUR O L,CALLET L,BARBA D,et al. A coherent computational approach to model bottom-up visual attention[J]. IEEE Transactions on Pattern Analysis&Ma-chine Intelligence,2006,28(5):802-817.
[3] GUO C,MA Q,ZHANG L. Spatio-temporal saliency detection using phase spectrum of quaternion Fourier transform[C]∥Proceedings of 2008 IEEE Conference on Computer Vision and Pattern Recognition. Anchorage:IEEE,2008:1-8.
[4] ITTI L,KOCH C,NIEBUR E. A model of saliency-based visual attention for rapid scene analysis[J]. IEEE Transactions on Pattern Analysis&Machine Intelligence,1998,20(11):1254-1259.
[5] ZHANG L,TONG M H,MARKS T K,et al. SUN:a Bayesian framework for saliency using natural statistics[J]. Journal of Vision,2008,8(7):32. 1-32. 20.
[6] HOU X D,ZHANG L Q. Saliency detection:a spectral residual approach[C]∥Proceedings of 2007 IEEE Conference on Computer Vision and Pattern Recognition. Minneapolis:IEEE,2007:1-8.
[7] LI J,LEVINE M D,AN X,et al. Visual saliency based on scale-space analysis in the frequency domain[J]. IEEE Transactions on Pattern Analysis&Machine Intelligence,2013,35(4):996-1010.
[8] JIANG B W,ZHANG L H,LU H C,et al. Saliency detection via absorbing Markov chain[C]∥Proceedings of2013 IEEE International Conference on Computer Vision.Sydney:IEEE,2013:1665-1672.
[9] SUN J,LU H,LIU X. Saliency region detection based on Markov absorption probabilities[J]. IEEE Transactions on Image Processing,2015,24(5):1639-1649.
[10] CHEN S,ZHENG L,HU X,et al. Discriminative saliency propagation with sink points[J]. Pattern Recognition,2016,60:2-12.
[11] SINGH A K,SAXENA R. On convolution and product theorems for FRFT[J]. Wireless Personal Communications,2012,65(1):189-201.
[12] WANG P G,TIAN H,ZHENG W. A novel image fusion method based on FRFT-NSCT[J]. Mathematical Problems in Engineering,2013,2013(11):408232/1-9.
[13] GAO L,QI L,WANG Y J,et al. Rotation invariance in2D-FRFT with application to digital image watermarking[J]. Journal of Signal Processing Systems,2013,72(2):133-148.
[14] WU J H,LUO X Z,ZHOU N R. Four-image encryption method based on spectrum truncation,chaos and the MODFrFT[J]. Optics&Laser Technology,2013,45(1):571-577.
[15] WANG X,QI L,TIE Y,et al. Face recognition based on weighted multi-order feature fusion of 2D-FrFT[M]∥Digital TV and Wireless Multimedia Communication:the13th International Forum. Shanghai:Springer,2016:137-145.