Multifocus image fusion scheme based on features of multiscale products and PCNN in lifting stationary wavelet domain

Multifocus image fusion aims at overcoming imaging cameras's finite depth of field by combining information from multiple images with the same scene. For the fusion problem of the multifocus image of the same scene, a novel algorithm is proposed based on multiscale products of the lifting stationary wavelet transform (LSWT) and the improved pulse coupled neural network (PCNN), where the linking strength of each neuron can be chosen adaptively. In order to select the coefficients of the fused image properly with the source multifocus images in a noisy environment, the selection principles of the low frequency subband coefficients and bandpass subband coefficients are discussed, respectively. For choosing the low frequency subband coefficients, a new sum modified-Laplacian (NSML) of the low frequency subband, which can effectively represent the salient features and sharp boundaries of the image in the LSWT domain, is an input to motivate the PCNN neurons; when choosing the high frequency subband coefficients, a novel local neighborhood sum of Laplacian of multiscale products is developed and taken as one type of feature of high frequency to motivate the PCNN neurons. The coefficients in the LSWT domain with large firing times are selected as coefficients of the fused image. Experimental results demonstrate that the proposed fusion approach outperforms the traditional discrete wavelet transform (DWT)-based, LSWT-based and LSWT-PCNN-based image fusion methods even though the source image is in a noisy environment in terms of both visual quality and objective evaluation.     

A multi-focus image fusion algorithm based on an improved dual-channel PCNN in NSCT domain

This paper presents a multi-focus image fusion algorithm based on dual-channel PCNN in NSCT domain. The fusion algorithm based on multi-scale transform is likely to produce the pseudo-Gibbs effects and it is not effective to fuse the dim or partial bright images. To solve these problems, this algorithm will get a number of different frequency sub-image of the two images by using the NSCT transform, the selection principles of different subband coefficients obtained by the NSCT decomposition are discussed in detail, and the images are fused based on the improved dual-channel PCNN in order to determine the band-pass sub-band coefficient, at last fused image is obtained by using the inverse NSCT transform. Fusion rules based on dual-channel PCNN are used to solve the complexity of the PCNN parameter settings and long computing time problems. The experimental results show that the algorithm has overcome the defects of the traditional multi-focus image fusion algorithm and improved the fusion effect.     

Multi-focus image fusion algorithm based on compound PCNN in Surfacelet domain

In order to effectively retain details and suppress noise, a multi-focus image fusion method based on Surfacelet transform and compound PCNN is proposed. Surfacelet transform is a powerful multi-resolution analysis tool which is able to decompose the original image into a number of different frequency band sub-images, compound PCNN model is a combined model of PCNN and dual-channel PCNN which is to select the fusion coefficients from the decomposed coefficients, the Local sum-modified-Laplacian (LSML) is selected as external stimulus of compound PCNN, fusion coefficients are decided by compound PCNN. The experimental results show that the new method has a good performance, fusion image has more texture details and it is more similar to the original images, the objective evaluation indexes show that this method is superior to the traditional image fusion methods.     

A medical image fusion method based on energy classification of BEMD components

A medical image fusion method based on bi-dimensional empirical mode decomposition (BEMD) and dual-channel PCNN is proposed in this paper. The multi-modality medical images are decomposed into intrinsic mode function (IMF) components and a residue component. IMF components are divided into high-frequency and low-frequency components based on the component energy. Fusion coefficients are achieved by the following fusion rule: high frequency components and the residue component are superimposed to get more textures; low frequency components contain more details of the source image which are input into dual-channel PCNN to select fusion coefficients, the fused medical image is achieved by inverse transformation of BEMD. BEMD is a self-adaptive tool for analyzing nonlinear and non-stationary data; it doesn’t need to predefine filter or basis function. Dual-channel PCNN reduces the computational complexity and has a good ability in selecting fusion coefficients. A combined application of BEMD and dual-channel PCNN can extract the details of the image information more effectively. The experimental result shows the proposed algorithm gets better fusion result and has more advantages comparing with traditional fusion algorithms.     

Multi-focus image fusion based on robust principal component analysis and pulse-coupled neural network

Multi-focus image fusion combines multiple source images with different focus points into one image, so that the resulting image appears all in-focus. In order to improve the accuracy of focused region detection and fusion quality, a novel multi-focus image fusion scheme based on robust principal component analysis (RPCA) and pulse-coupled neural network (PCNN) is proposed. In this method, registered source images are decomposed into principal component matrices and sparse matrices with RPCA decomposition. The local sparse features computed from the sparse matrix construct a composite feature space to represent the important information from the source images, which become inputs to PCNN to motivate the PCNN neurons. The focused regions of the source images are detected by the firing maps of PCNN and are integrated to construct the final, fused image. Experimental results demonstrate that the superiority of the proposed scheme over existing methods and highlight the expediency and suitability of the proposed method.     

A novel image fusion algorithm based on nonsubsampled shearlet transform

To overcome the shortcoming of traditional image fusion method based on multi-scale transform, a novel adaptive image fusion algorithm based on nonsubsampled shearlet transform (NSST) is proposed. Firstly, the NSST is utilized to decompose the source images on various scales and in different directions, and the low frequency sub-band and bandpass sub-band coefficients are obtained. Secondly, for the low frequency sub-band coefficients, the singular value decomposition method in the gradient domain is used to estimate the local structure information of image, and an adaptive ‘weighted averaging’ fusion rule based on the sigmoid function and the extracted features is presented. To improve the quality of fused image, a novel sum-modified-Laplacian (NSML), which can extract more useful information from source images, is employed as the measurement to select bandpass sub-band coefficients. Finally, the fused image is obtained by performing the inverse NSST on the combined coefficients. The proposed fusion method is verified on several sets of multi-source images, and the experimental results show that the proposed approach can significantly outperform the conventional image fusion methods in terms of both objective evaluation criteria and visual quality.     

基于特征提取和脉冲耦合神经网络的图像融合算法

为了提高目标检测的准确度和精确度,提出了一种新的基于特征提取和脉冲耦合神经网络(Pulse-coupled neural network,PCNN)模型的融合算法。该算法的原理是:首先对原图像进行特征提取,然后分别建立基于处理后的图像及原图像的PCNN。融合的作用是通过PCNN的链接输入得以体现的。融合图像综合了各个侧重于原图像不同特征的滤波器的选择性。仿真结果证明,所提出的算法在增加图像目标的可识别程度方面效果明显。     

Image fusion algorithm based on redundant-lifting NSWMDA and adaptive PCNN

Pyramid decomposition in the NSCT transformation is a band-pass filtering process in the frequency domain where different scales of images are orthogonal. However, from the perspective of the image content, correlation is likely to exist between the fused images, and this kind of decomposition makes images of different scales contain redundant information, as a result of which the fused image may not capture the subtle information from the original images. In order to overcome the above-mentioned problem, an effective image fusion method based on redundant-lifting non-separable wavelet multi-directional analysis (NSWMDA) and adaptive pulse coupled neural network (PCNN) has been proposed. The original images are firstly decomposed by using the NSWMDA into several sub-bands in order to retain texture detail and contrast information of the images, and then adaptive PCNN algorithm is applied on the high-frequency directional sub-bands to extract the high-frequency information. The low-frequency sub-bands are evaluated by weighted average based on Gaussian kernel with a chosen maximum fusion rule. Results from experiments show that the proposed method can make the fused image maintains more texture details and contrast information.     

基于平移不变剪切波变换域图像融合算法

针对传统基于多尺度变换的图像融合方法存在的缺点,提出了一种基于平移不变剪切波变换域的自适应图像融合新方法.首先,使用平移不变剪切波变换对源图像进行分解,得到低频子带及方向带通子带系数.然后,对于低频子带系数采用梯度域奇异值分解方法估计图像的局部结构信息,提出了基于提取的特征与S函数的可变加权融合策略;对于各方向带通子带系数,提出了一种基于改进的拉普拉斯能量和匹配的"加权平均"和选择相结合的系数选择策略.最后,对得到的融合系数进行逆变换得到融合图像.通过实验可以发现相比于传统的图像融合方法,本文方法得到了更高的客观指标,融合图像视觉效果更好.     

有限离散剪切波域的红外可见光图像融合

针对目前图像融合过程中的不足之处,结合有限离散剪切波具有高的方向敏感性和抛物尺度化特性,提出了一种有限离散剪切波变换下的图像融合算法。首先对严格配准的多传感器图像进行有限离散剪切波变换,得到低频子带系数和不同尺度不同方向的高频子带系数;然后对低频子带系数采用全局特征值和像素点之间的差异性与区域空间频率匹配度相结合的融合算法,高频方向子带系数采用方向权重对比度与相对区域平均梯度和相对区域方差相结合的方案;最后通过有限离散剪切波逆变换得到融合图像。实验结果表明,与其他的融合算法相比较,本文算法不但有良好的主观视觉效果,而且3幅图像的客观评价指标分别平均提高了0.9%、3.8%、3.1%,2.6%、3.8%、2.9%和1.5%、125%、59%,充分说明了本文融合算法的优越性。     

Technique for gray-scale visual light and infrared image fusion based on non-subsampled shearlet transform

A novel image fusion technique based on NSST (non-subsampled shearlet transform) is presented, aiming at resolving the fusion problem of spatially gray-scale visual light and infrared images. NSST, as a new member of MGA (multi-scale geometric analysis) tools, possesses not only flexible direction features and optimal shift-invariance, but much better fusion performance and lower computational costs compared with several current popular MGA tools such as NSCT (non-subsampled contourlet transform). We specifically propose new rules for the fusion of low and high frequency sub-band coefficients of source images in the second step of the NSST-based image fusion algorithm. First, the source images are decomposed into different scales and directions using NSST. Then, the model of region average energy (RAE) is proposed and adopted to fuse the low frequency sub-band coefficients of the gray-scale visual light and infrared images. Third, the model of local directional contrast (LDC) is given and utilized to fuse the corresponding high frequency sub-band coefficients. Finally, the final fused image is obtained by using inverse NSST to all fused sub-images. In order to verify the effectiveness of the proposed technique, several current popular ones are compared over three different publicly available image sets using four evaluation metrics, and the experimental results demonstrate that the proposed technique performs better in both subjective and objective qualities.     

基于小波包变换的红外与可见光图像融合

针对红外与可见光图像特点,提出一种基于小波包变换的融合算法。该算法先对源图像进行小波包分解,得到低频分量和各带通方向子带分量,并对不同分量采用不同的融合规则进行融合处理,得到各融合系数,然后经小波包重构获得融合图像。该方法可提取源图像细节信息,取得较好的融合效果。     

基于变换域PCNN的近红外与彩色可见光融合

针对近红外与彩色可见光图像融合后对比度低、细节丢失和颜色失真等问题,提出一种基于多尺度变换和自适应脉冲耦合神经网络(PCNN-pulse coupled neural network,PCNN)的红外与彩色可见光图像融合的新算法。首先将彩色可见光图像转换到HSI（hue saturation intensity)空间,HSI色彩空间包含亮度、色度和饱和度三个分量,并且这三个分量互不相关,因此利用这个特点可对三个分量分别进行处理。将其亮度分量与近红外图像分别进行多尺度变换,变换方法选择Tetrolet变换。变换后分别得到低频和高频分量,针对图像低频分量,提出一种期望最大的低频分量融合规则;针对图像高频分量,采用高斯差分算子调节PCNN模型的阈值,提出一种自适应的PCNN模型作为融合规则。处理后的高低频分量经过Tetrolet逆变换得到的融合图像作为新的亮度图像。然后将新的亮度图像和原始的色度和饱和度分量反向映射到RGB空间,得到融合后的彩色图像。为了解决融合带来的图像平滑化和原始图像光照不均的问题,引入颜色与锐度校正机制(colour and sharpness correction, CSC)来提高融合图像的质量。为了验证方法的有效性,选取了5组分辨率为1 024×680近红外与彩色可见光图像进行试验,并与当前高效的四种融合方法以及未进行颜色校正的本方法进行了对比。实验结果表明,同其他图像融合算法进行对比分析,该方法在有无CSC颜色的情况下均能保留最多的细节和纹理,可见度均大大提高,同时本方法的结果在光照条件较弱的情况下具有更多的细节和纹理,均具有更好的对比度和良好的色彩再现性。在信息保留度、颜色恢复、图像对比度和结构相似性等客观指标上均具有较大优势。     

基于可见光的多波段偏振图像融合新算法

采用了一种新的基于小波变换的偏振图像融合算法.首先,将两个波段中的每一波段三幅偏振图像利用小波变换分解成低频和高频部分,低频的小波系数平均值作为融合后的低频系数,高频细节系数根据不同区域特征选择方法以及对应输入图像小波系数的窗口区域方差来确定融合后高频小波系数,得到一个波段一幅图像.接着,将得到的图像再进行小波分解,采用低频图像的小波系数最小值作为融合后的低频系数,高频图像根据纹理一致性测度的纹理检测确定融合规则,用来调整高频小波系数,将来自不同图像的特征与细节融合在一起,并对融合图像质量进行了对比评价.实验结果表明,融合后的偏振图像不仅反映了场景的偏振信息,而且还包含了丰富的光谱信息,目标与背景的衬比度也得到了增强,为进一步的目标检测和识别提供了便利.     

改进的曲波变换图像融合方法

考虑将曲波变换引入图像融合能够更好地提取原始图像,对一种新的图像融合方法—曲波变换图像融合法进行了研究。将图像序列进行曲波变换后,通过对所有图像的高频进行逆变换及域值处理来获得区域图。根据区域图中高频区域的边界点在每张图层上的活跃度不同求得区域边界的图层分布,利用插值获得高频区域的区域分布图。通过高频区域的膨胀求得整幅图的区域分布图,然后在曲波变换的变换域,利用区域分布图对多尺度的高频系数采用高斯加权求和;对低频系数采用取平均值的规则完成图像的融合。进行了图像融合实验,实验结果表明,与传统的小波变换及基于像素的曲波变换相比,提出的方法获得的融合图像边缘更清晰,更接近参考图像。     

Global and local contrast enhancement algorithm for image using wavelet neural network and stationary wavelet transform

A new contrast enhancement algorithm for image is proposed employing wavelet neural network (WNN)and stationary wavelet transform (SWT). Incomplete Beta transform (IBT) is used to enhance the global contrast for image. In order to avoid the expensive time for traditional contrast enhancement algorithms,which search optimal gray transform parameters in the whole gray transform parameter space, a new criterion is proposed with gray level histogram. Contrast type for original image is determined employing the new criterion. Gray transform parameter space is given respectively according to different contrast types,which shrinks the parameter space greatly. Nonlinear transform parameters are searched by simulated annealing algorithm (SA) so as to obtain optimal gray transform parameters. Thus the searching direction and selection of initial values of simulated annealing is guided by the new parameter space. In order to calculate IBT in the whole image, a kind of WNN is proposed to approximate the IBT. Having enhanced the global contrast to input image, discrete SWT is done to the image which has been processed by previous global enhancement method, local contrast enhancement is implemented by a kind of nonlinear operator in the high frequency sub-band images of each decomposition level respectively. Experimental results show that the new algorithm is able to adaptively enhance the global contrast for the original image while it also extrudes the detail of the targets in the original image well. The computation complexity for the new algorithm is O(MN) log(MN), where M and N are width and height of the original image, respectively.     

一种基于图像融合和卷积神经网络的相位恢复方法

相位恢复法利用光波传输中某一(或某些)截面上的光强分布来传感系统波前,其结构简单,不易受震动及环境干扰,被广泛应用于光学遥感和像差检测等领域.传统相位恢复法采用迭代计算,很难满足实时性要求,且在一定程度上依赖于迭代转换或迭代优化初值.为克服上述问题,本文提出了一种基于卷积神经网络的相位恢复方法,该方法采用基于小波变换的图像融合技术对焦面和离焦面图像进行融合处理,可在不损失图像信息的同时简化卷积神经网络的输入.网络模型训练完成后可依据输入的融合图像直接输出表征波前相位的4-9阶Zernike系数,且波前传感精度均方根(root-mean-square,RMS)可达0.015λ,λ=632.8 nm.研究了噪声、离焦量误差和图像采样分辨率等因素对波前传感精度的影响,验证了该方法对噪声具有一定鲁棒性,相对离焦量误差在7.5%内时,波前传感精度RMS仍可达0.05λ,且随着图像采样分辨率的提升,波前传感精度有所改善,但训练时间成本随之增加.此外,分析了实际应用中,当系统像差阶数与网络训练阶数略有差异时,本方法所能实现的传感精度,并给出了解决方案.     

潜在低秩表示框架下基于卷积神经网络结合引导滤波的红外与可见光图像融合

为提高融合图像的可视性,解决传统红外与可见光图像融合算法中存在的边缘特征缺失、细节模糊的问题,提出了一种潜在低秩表示框架下基于卷积神经网络结合引导滤波的红外与可见光图像融合算法。该算法首先利用潜在低秩表示对源图像进行分解,得到源图像的低秩分量和显著分量。其次,利用卷积神经网络根据源图像的特征信息,得到权值图。再次,通过引导滤波算法对权值图进行边缘锐化,然后再将优化后的权值图分别与源图像的低秩分量和显著分量融合,得到融合图像的低秩分量和显著分量。最后,将融合图像的低秩分量和显著分量叠加,得到最终的融合图像。实验结果表明,该算法在主观评价和客观指标上均优于传统的红外与可见光图像融合算法。     

Novel fusion method for visible light and infrared images based on NSST–SF–PCNN

The purpose of image fusion is to combine the source images of the same scene into a single composite image with more useful information and much better visual effects, which is undoubtedly suitable for further image processing tasks. This paper presented a novel fusion method for visible light and infrared images based on non-subsampled shearlet transform (NSST)–spatial frequency (SF)–pulse coupled neural network (PCNN). As a recently developed multi-resolution geometric analysis tool, NSST not only has remarked superiorities over other past conventional tools in terms of information capturing and computational costs saving, but also overcomes the lack of shift-invariance in shearlet transform (ST), so NSST applies to conducting the decompositions and reconstructions. Besides, traditional PCNN model is also upgraded to be an improved one called IPCNN in this paper to fuse the low-frequency and high-frequency subband coefficients. In the IPCNN structure, on the one hand, the value of the linking strength β is determined by the SF which represents the gradient features of the subband image; on the other hand, the time matrix is utilized to adaptively decide the iteration number of the IPCNN model, which is helpful to increase the function efficiency and save computational resources. Experimental results indicate that the proposed method performs well and has obvious superiorities over other current typical ones in both subjective visual performance and objective criteria.     

基于HVS的小波变换数字图像水印算法

为实现图像水印在数据认证、访问控制和版权保护等领域发挥作用,提出一种人眼视觉感知特性的小波变换数字图像水印算法。首先对二值图像水印进行Arnold变换,并采用行堆叠的方法,转换成一维序列,然后根据密钥生成的高斯序列中值的正负来确定水印信息嵌入时是采用那种调制方式,最后对原始图像进行3级DWT分解,在小波图像的第二层中频子带中,结合人眼视觉特性将水印正向或负向调制叠加在宿主图像大于JND 值的小波系数上。实验结果表明:该算法能抵抗JPEG压缩、加噪、剪切等多种攻击,具有良好的透明性,鲁棒性和安全性。