The discrete wavelet transform has become an attractive tool for fusing multisensor images. This paper investigates the discrete wavelet frame transform. A major advantage of this method over discrete wavelet transfor...The discrete wavelet transform has become an attractive tool for fusing multisensor images. This paper investigates the discrete wavelet frame transform. A major advantage of this method over discrete wavelet transform is aliasing free and translation invariant. The discrete wavelet frame (DWF) transform is used to decompose the registered images into multiscale representation with the low frequency and the high frequency bands. The low frequency band is normalized and fused by using the expectation maximization (EM) algorithm. The informative importance measure is applied to the high frequency band. The final fused image is obtained by taking the inverse transform on the composite coefficient representations. Experiments show that the proposed method is more effective than conventional image fusion methods.展开更多
为了在模型参数先验分布知识未知情况下实现基于区域和统计的图像分割,并同时获取更加精确的模型参数,提出了一种结合Voronoi划分技术、最大期望值(Expectation Maximization,EM)和最大边缘概率(Maximizationof the Posterior Marginal,...为了在模型参数先验分布知识未知情况下实现基于区域和统计的图像分割,并同时获取更加精确的模型参数,提出了一种结合Voronoi划分技术、最大期望值(Expectation Maximization,EM)和最大边缘概率(Maximizationof the Posterior Marginal,MPM)算法的图像分割方法。该方法利用Voronoi划分技术将图像域划分为若干子区域,待分割图像中的同质区域可以由一组子区域拟合而成,并假定各同质区域内像素强度服从同一独立的正态分布,从而建立图像模型,然后结合EM/MPM算法进行图像分割和模型参数估计,其中,MPM算法用于实现面向同质区域的图像分割,EM算法用于估计图像模型参数。为了验证提出的图像分割方法,分别对合成图像和真实图像进行了分割实验,并和传统的基于像素的MRF分割结果进行对比,测试结果的定性和定量分析表明了该方法的有效性和准确性。展开更多
基金国家自然科学基金，Shanghai Key Scientific Project，China Ph.D.Discipline Special Foundation
文摘The discrete wavelet transform has become an attractive tool for fusing multisensor images. This paper investigates the discrete wavelet frame transform. A major advantage of this method over discrete wavelet transform is aliasing free and translation invariant. The discrete wavelet frame (DWF) transform is used to decompose the registered images into multiscale representation with the low frequency and the high frequency bands. The low frequency band is normalized and fused by using the expectation maximization (EM) algorithm. The informative importance measure is applied to the high frequency band. The final fused image is obtained by taking the inverse transform on the composite coefficient representations. Experiments show that the proposed method is more effective than conventional image fusion methods.
文摘为了在模型参数先验分布知识未知情况下实现基于区域和统计的图像分割,并同时获取更加精确的模型参数,提出了一种结合Voronoi划分技术、最大期望值(Expectation Maximization,EM)和最大边缘概率(Maximizationof the Posterior Marginal,MPM)算法的图像分割方法。该方法利用Voronoi划分技术将图像域划分为若干子区域,待分割图像中的同质区域可以由一组子区域拟合而成,并假定各同质区域内像素强度服从同一独立的正态分布,从而建立图像模型,然后结合EM/MPM算法进行图像分割和模型参数估计,其中,MPM算法用于实现面向同质区域的图像分割,EM算法用于估计图像模型参数。为了验证提出的图像分割方法,分别对合成图像和真实图像进行了分割实验,并和传统的基于像素的MRF分割结果进行对比,测试结果的定性和定量分析表明了该方法的有效性和准确性。
