摘要
提出一种基于稀疏表示和脉冲耦合神经网络(pulse coupled neural network,PCNN)的新方法。首先将原图像进行bandelet变换,提取出图像中的几何流和bandelet系数等重要信息,再利用PCNN进行几何流融合、根据稀疏相似度优化融合后的几何流,然后更新部分bandelet系数并根据最大绝对值规则进行融合,最后通过bandelet逆变换得到融合后的图像。仿真实验结果表明,本算法有效改善了融合效果,融合图像边缘、纹理清晰,整体效果极佳;与现有的平均值融合算法、拉普拉斯金字塔算法以及基于小波变换和PCNN的WT-PCNN算法相比,本算法得到的融合图像的灰度均值、标准差、平均梯度、互信息等指标都得到了提高。
A novel algorithm for image fusion was proposed based on sparse representation and PCNN(pulse coupled neural network).The bandelet transform was used to extract important information such as geometric flows and bandelet coefficients of the source image.Then geometric flows were fused by PCNN and optimized according to similarity of sparseness.Then,the bandelet coefficients were updated and fused according to a rule of maximum absolute.Finally,the inverse bandelet transform was applied for the fused image.The experimental results showed that this algorithm could effectively improve the fusion effect.The fusion image had clear edges,texture and excellent overall effect.Compared with the average algorithm,the Laplace pyramid algorithm and the WT-PCNN algorithm based on wavelet transform and PCNN,a proposed algorithm achieved the better average gray,standard deviation,average gradient and mutual information.
引文
[1]DAVID L H.An introduction to multisensor data fusion[J].Proceedings of the IEEE,1997,85(1):6-23.
[2]SIMONE G,FARINA A,MORABITO F C,et al.Image fusion techniques for remote sensing applications[J].Information Fusion,2002,3(1):3-15.
[3]QU Guihong,ZHANG Dali,YAN Ping.Medical image fusion by wavelet transform modulus maxima[J].Optics Express,2001,9(4):184-190.
[4]LI Shutao,YANG Bin.Multifocus image fusion by combining curvelet and wavelet transform[J].Pattern Recognition Letters,2008,29(9):1295-1301.
[5]BURT P J,ADELSON E H.The Laplacian pyramid as a compact image code[J].IEEE Transactions on Communications,1983,31(4):532-540.
[6]CANDS E,DONOHO D.Curvelets:a surprisingly effective nonadaptive representation of objects with edges[R].Nashville:Vanderbilt University Press,1999.
[7]DO M N,VETTERLI M.The contourlet transform:an efficient directional multiresolution image representation[J].IEEE Transactions on Image Processing,2005,14(12):2091-2106.
[8]PENNEC E LE,MALLAT S.Sparse geometric image representations with bandelets[J].IEEE Trans on Image Processing,2005,14(4):423-438.
[9]DONOHO D.Wedgelets:nearly-minimax estimation of edges[J].Annals of Statistics,1999,27(3):859-897.
[10]ECKHORN R,REITBOECK H,ARNDT M,et al.Feature linking via synchronization among distributed assemblies:Simulations of results from cat visual cortex[J].Neural Computation,1990,2(3):293-307.
[11]PADGETT M L,JOHNSON J L.Pulse coupled neural networks(PCNN)and wavelets:biosensor applications[C]//International Conference on Neural Networks,Piscataway:IEEE,1997:2507-2512.
[12]KUNTIMAD G,RANGANATH H S.Perfect image segmentation using pulse coupled neural networks[J].Neural Networks,1999,10(3):591-598.
[13]JOHNSON J L,PADGETT M L.PCNN models and applications[J].IEEE Transactions on Neural Networks,1999,10(3):480-498.
[14]YAMAGUCHI Y,ISHIMUR A K,WADA M.Synchronized oscillation and dynamical clustering in chaotic PCNN[C]//Proceedings of the 41st SICE Annual Conference.[S.l.]:Society of Instrument and Control Engineers,2002:730-735.
[15]MA Yide,LIU Qing,QIAN Zhibai.Automated image segmentation using improved PCNN model based on cross-entropy[C]//Proceedings of 2004 International Symposium on Intelligent Multimedia,Video and Speech Proccessing.Piscataway,IEEE:2004:743-746.
[16]GU Xiaodong,ZHANG Liming.Orientation detection and attention selection based unit-linking PCNN[C]//International Conference on Neural Networks&Brain,Piscataway:IEEE,2005:1328-1333.
[17]LI Wei,ZHU Xuefeng.A new image fusion algorithm based on wavelet packet analysis and PCNN[C]//International Conference on Machine Learning and Cybernetics.Piscataway:IEEE,2005:5297-5301.
[18]JOHNSON J L,PADGETT M L.PCNN models and applications[J].IEEE Transactions on Neural Networks,1999,10(3):480-498.
[19]BURT P,ADELSON E.The laplacian pyramid as a compact Image code[J].IEEE Transactions on Communications,1983,31(4):532-540.