一种基于区域相似性的图像融合评价方法

本文提出一种新的图像融合客观评价方法.首先,利用模糊C-均值聚类算法（FCM）对图像进行区域分割,获取区域特征矩阵.通过降维处理,计算区域的距离,以此作为区域相似度.其次,以对应区域的特征向量元素相似比作为对应区域的权重.区域的边缘强度比值作为区域之间的权重.最终获得图像的相似度.实验结果表明这种相似性度量考虑了图像像素的局部关系以及区域的显著性,更加符合人类的视觉特征.     

基于区域最大相似度的快速图像分割算法

针对基于区域最大相似度图像分割(MSRM)算法中利 用颜色直方图描述符计算相邻区域之间相似度存 在着计算量大和描述能力的不足,提出改用基于局部二值模式(LBP)纹理描述符计 算相邻区域相似度的改进MSRM(IMSRM)算法。LBP描述符通过对像素点之间局 部微结构进行编码实 现了对其空间关系的描述,有效提高了对区域特征的描述能力,并且所获得的特征矢量维数 远小于颜色直 方图,区域之间的相似度计算效率大幅度得到提高。与MSRM算法对比实验表明,IMSRM算法 可以从各种复杂背景中有效提取待分割对象的轮廓,所提取的轮廓边缘细节上更优,算法执 行效率能够提高50%左右。     

基于引导滤波及视差图融合的立体匹配

立体匹配是双目视觉领域的重要研究方向.为在保证图片纹理区域匹配精度的同时降低弱纹理区域的误匹配率,提出一种基于引导滤波及视差图融合的立体匹配方法.首先,根据图像颜色相似性将图片划分为纹理较丰富区域和弱纹理区域.接着,分别采用不同参数的引导滤波进行代价聚合及视差计算,得到两张视差图.然后依据纹理区域划分的结果对获得的两张视差图进行融合.最后,通过左右一致性检测、加权中值滤波等视差优化步骤得到最终视差图.对Middlebury测试平台上标准图像对的实验结果表明,该方法在6组弱纹理图像上的平均误匹配率为9.67％,较传统引导滤波立体匹配算法具有更高的匹配精度.     

The dual-bootstrap iterative closest point algorithm with application to retinal image registration

Motivated by the problem of retinal image registration, this paper introduces and analyzes a new registration algorithm called Dual-Bootstrap Iterative Closest Point (Dual-Bootstrap ICP). The approach is to start from one or more initial, low-order estimates that are only accurate in small image regions, called bootstrap regions. In each bootstrap region, the algorithm iteratively: 1) refines the transformation estimate using constraints only from within the bootstrap region; 2) expands the bootstrap region; and 3) tests to see if a higher order transformation model can be used, stopping when the region expands to cover the overlap between images. Steps 1): and 3), the bootstrap steps, are governed by the covariance matrix of the estimated transformation. Estimation refinement [Step 2)] uses a novel robust version of the ICP algorithm. In registering retinal image pairs, Dual-Bootstrap ICP is initialized by automatically matching individual vascular landmarks, and it aligns images based on detected blood vessel centerlines. The resulting quadratic transformations are accurate to less than a pixel. On tests involving approximately 6000 image pairs, it successfully registered 99.5% of the pairs containing at least one common landmark, and 100% of the pairs containing at least one common landmark and at least 35% image overlap.     

基于交互式分割和勾画匹配的颜色迁移

针对经典全局颜色迁移算法在处理具有复杂色彩的自然图像时常出现较多颜色误传的问题,提出一种基于交互式分割与勾画匹配的局部自适应颜色迁移算法。该算法不仅适用于彩色图像间的颜色迁移,而且适用于灰度图着色。算法基本思路是首先分别对两幅图像进行交互式分割,然后对分割后图像的同质景物区域进行勾画匹配,最后在相匹配区域间进行局部颜色迁移。实验结果表明该算法能很好地实现参考图像和目标图像中同质景物内容区域间的颜色迁移,减少颜色误传的现象。结果图像颜色自然度与局部景物一致度有明显提高。     

基于深度哈希算法的极光图像分类与检索方法

面对形态万千、变化复杂的海量极光数据,对其进行分类与检索为进一步研究地球磁场物理机制和空间信息具有重要意义。该文基于卷积神经网络(CNN)对图像特征提取方面的良好表现,以及哈希编码可以满足大规模图像检索对检索时间的要求,提出一种端到端的深度哈希算法用于极光图像分类与检索。首先在CNN中嵌入空间金字塔池化(SPP)和幂均值变换(PMT)来提取图像中多种尺度的区域信息;其次在全连接层之间加入哈希层,将全连接层最能表现图像的高维语义信息映射为紧凑的二值哈希码,并在低维空间使用汉明距离对图像对之间的相似性进行度量;最后引入多任务学习机制,充分利用图像标签信息和图像对之间的相似度信息来设计损失函数,联合分类层和哈希层的损失作为优化目标,使哈希码之间可以保持更好的语义相似性,有效提升了检索性能。在极光数据集和 CIFAR-10 数据集上的实验结果表明,所提出方法检索性能优于其他现有检索方法,同时能够有效用于极光图像分类。     

基于多目标区域的图像检索

针对传统的基于目标区域的图像检索算法中存在的"语义鸿沟"问题,以及基于全局特征的图像检索算法不能很好地处理多目标检索问题,提出了一种基于多目标区域的图像检索模型,并实现了一款高效的检索算法.首先借助于目标检测算法定位出图像中的目标,然后使用卷积神经网络(CNN)提取各个目标的特征,最后采用新提出的多目标区域相似度测量方法计算其与数据库图像的相似度并返回检索结果.实验表明,所提算法与现有的其他检索算法相比,在多目标图像检索任务上性能更佳.     

基于KL散度及多尺度融合的显著性区域检测算法

基于对超像素颜色概率分布间KL散度的计算,以及对多尺度显著图的融合处理,该文提出一种新的显著性区域检测算法。首先,采用超像素算法多尺度分割图像,在各尺度下用分割产生的超像素为节点,并依据超像素分割数量对各超像素进行适当邻接连通扩展,构建无向扩展闭环连通图。 其次,依据颜色判别力聚类量化各超像素内颜色,统计颜色聚类标签的概率分布,用概率分布间KL散度的调和平均值为扩展闭环连通图的边加权,再依据区域对比度并结合边界连通性,获取各尺度下的显著图。 最后,平均融合各尺度下显著图,并进行优化处理,得到最终的显著图。 在一些大型参考数据集上进行大量实验表明,所提算法优于当前一些先进算法,具有较高精确度和召回率,并且可以产生平滑显著图。     

Wavelet-based level set evolution for classification of textured images

We present a supervised classification model based on a variational approach. This model is specifically devoted to textured images. We want to get a partition of an image, composed of texture regions separated by regular interfaces. Each kind of texture defines a class. We use a wavelet packet transform to analyze the textures, characterized by their energy distribution in each sub-band. In order to have an image segmentation according to the classes, we model the regions and their interfaces by level set functions. We define a functional on these level sets whose minimizers define the optimal classification according to texture. A system of coupled PDEs is deduced from the functional. By solving this system, each region evolves according to its wavelet coefficients and interacts with the neighbor regions in order to obtain a partition with regular contours. Experiments are shown on synthetic and real images.     

Copy-move image forgery detection based on Gabor magnitude

With advancement of media editing software, even people who are not image processing experts can easily alter digital images. Various methods of digital image forgery exist, such as image splicing, copy-move forgery, and image retouching. The most common method of tampering with a digital image is copy-move forgery, in which a part of an image is duplicated and used to substitute another part of the same image at a different location. In this paper, we present an efficient and robust method to detect such artifacts. First, the tampered image is segmented into overlapping fixed-size blocks, and the Gabor filter is applied to each block. Thus, the image of Gabor magnitude represents each block. Secondly, statistical features are extracted from the histogram of orientated Gabor magnitude (HOGM) of overlapping blocks, and reduced features are generated for similarity measurement. Finally, feature vectors are sorted lexicographically, and duplicated image blocks are identified by finding similarity block pairs after suitable post-processing. To enhance the algorithm’s robustness, a few parameters are proposed for removing the wrong similar blocks. Experiment results demonstrate the ability of the proposed method to detect multiple examples of copy-move forgery and precisely locate the duplicated regions, even when dealing with images distorted by slight rotation and scaling, JPEG compression, blurring, and brightness adjustment.     

基于小波模糊聚类区域分割的图像检索

基于内容的图像检索是近年来的研究热点,本文给出一种基于区域分割的图像检索算法。算法首先对图像按JPEG2000标准进行小波变换,对变换得到的低频子带提取一定的颜色和纹理特征用于模糊聚类,从而将图像的低频子带分割为一定的区域,将分割结果映射回整幅图像,提取整幅图像各个区域的特征矢量,用于区域相似度比较,最后按照一定的区域匹配准则得到整幅图像之间的相似度。实验结果表明,本文算法具有良好的图像检索性能。     

Location algorithm for seal imprints on Chinese bank-checks based on region growing

Aslegalconfirmationofbankcheck,sealimprintsare usedinOrientalcountries.Manymethodsonautomatic sealimprintverificationbyacomputerhavebeenan nounced[14],ofwhich,however,mostcannotbeapplied directly,becausetheydealwiththoseoftheimagesof onlyasealimprintother…     

A micro-image fusion algorithm based on region growing

Due to the limitation of Depth Of Field (DOF) of microscope, the regions which are not within the DOF will be blurring after imaging. Thus for micro-image fusion, the most important step is to identify the blurring regions within each micro-image, so as to remove their undesirable impacts on the fused image. In this paper, a fusion algorithm based on a novel region growing method is proposed for micro-image fusion. The local sharpness of micro-image is judged block by block, then blocks whose sharpness is lower than an adaptive threshold are used as seeds, and the sharpness of neighbors of each seed are evaluated again during the region growing until the blurring regions are identified completely. With the decreasing in block size, the obtained region segmentation becomes more and more accurate. Finally, the micro-images are fused with pixel-wise fusion rules. The experimental results show that the proposed algorithm benefits from the novel region segmentation and it is able to obtain fused micro-image with higher sharpness compared with some popular image fusion method.     

基于Wasserstein距离的局部能量分割模型

 提出了一种基于Wasserstein距离和图像局部区域直方图信息的非参数活动轮廓分割模型.用该距离对图像中不同区域的直方图进行比较,提高了相似性衡量的准确性;引入高斯内核函数来获取图像局部区域直方图信息,并将信息嵌入模型指导轮廓演化,以克服由于亮度不均造成的图像分割困难;通过水平集规范项提高计算精度并避免水平集演化的重新初始化.实验结果表明,本模型能够对亮度不均的无序特征图像进行有效准确的分割.     

Adaptive parametric estimation and classification of remotelysensed imagery using a pyramid structure

An unsupervised region-based image segmentation algorithm implemented with a pyramid structure has been developed. Rather than depending on traditional local splitting and merging of regions with a similarity test of region statistics, the algorithm identifies the homogeneous and boundary regions at each level of the pyramid; the global parameters of each class are then estimated and updated with the values of the homogeneous regions represented at that level of the pyramid using mixture distribution estimation. The image is then classified through the pyramid structure. Classification results obtained for both simulated and SPOT imagery are presented     

Quantisation error curve: a measure of similarity betweenrequantised images

A new similarity measure between a digital image and its quantised version is presented. By calculating the quantisation error on each of a sequence of subdomains, we can obtain a quantisation error curve as a similarity measure. This measure outperforms the commonly used measure for estimating the similarity between requantised images     

一种基于颜色连通的图像纹理检索新方法

提出并实现一种结合图像颜色连通区域信息及其纹理特征的图像检索新方法.首先提取图像的分块主颜色,根据提出的相关颜色定义,搜索确定图像中的颜色连通区域集.然后,提取图像中各颜色连通区域对应的四种颜色共生矩阵特征,利用针对该特征设计的图像相似性度量函数实现基于内容的图像检索.实验结果表明,该方法能有效结合图像的纹理信息及其颜色构成和分布信息,具有良好的检索效果和性能.