结合深度迭代缩放卷积神经网络的PRNU提取算法

光响应非均匀性（photo-response non-uniformity，PRNU）是用于数字图像设备溯源的一种重要特征，也被称为成像设备指纹。针对图像真实噪声包含PRNU和大量未知噪声的复杂特性，本文提出一种结合深度迭代缩放卷积神经网络的PRNU数字成像设备指纹提取算法。首先，通过连续重复的缩小与放大特征图的分辨率来提高GPU内存利用效率和生成大的感受野，尽可能的提取包含完整PRNU指纹的真实噪声。然后，利用来自同一数字成像设备多幅图像的噪声残差来估计PRNU指纹。本文算法在相机溯源数据集Dresden和手机溯源数据集Daxing上进行了测试。实验结果显示，与传统方法相比本文算法具有的更好的识别率和普适性。      

基于Blob-Harris特征区域和NSCT-Zernike的鲁棒水印算法

为了有效抵抗水印图像的几何攻击,该文提出了一种基于Blob-Harris特征区域和非下采样轮廓波变换(NSCT)和伪Zernike矩的鲁棒水印算法。首先原始图像进行两层非下采样Contourlet变换后提取其低频图像,然后利用Blob-Harris检测算子对低频图像进行特征点提取,根据各个特征点的特征尺度确定其特征区域,优化筛选出稳定且互不重叠的特征区域并将其四周补零,得到稳定的互不重叠的方形特征区域作为水印嵌入区域,最后计算每一个方形特征区域的Zernike矩,将水印信息嵌入在量化调制正则化Zernike矩的幅值当中。实验结果表明,Lena图峰值信噪比达到40 dB以上时,本文算法对常规图像处理以及缩放、旋转、剪切等几何攻击和组合攻击都有相对较强的鲁棒性。     

Automatic panorama with auto-focusing based on image fusion for microscopic imaging system

It is usually desirable from a microscope imaging system to have an efficient auto-focusing and to maintain imaging quality throughout microscopy screening restricted automatically by the specimen borders. This paper presents a novel image fusion-based auto-focusing method and an automatic panorama confined with surroundings of the specimen so as to minimize the auto-scanning time for microscope imaging system. Multi-focus color image fusion is proposed to achieve the auto-focusing task for microscopic imaging. An image sequence is captured by using a microscope eyepiece camera with moving the microscope stage along Z-axis. Several images around a reference image are used to achieve in-focus image, instead of selecting a single image from the sequence. The reference image is an image given highest focus measurement value within the image sequence. Moreover, various evaluation criteria are utilized to analyze the performance of the proposed auto-focus approach on different color models for microscopic imaging. Microscope stage position along the Z-axis is automatically adjusted by image processing-based feedback system to maintain focus during scanning process. In this screening, the in-focus images with overlapped areas on the X–Y axes are stitched together to produce a mosaic image without any seams. In this process, the screening area is automatically constrained with the specimen regions which occupy 20–40 % of the glass surface. An artificial neural network-based learning algorithm is implemented to decide whether the specimen regions are within microscope objective field of view or not. The experimental studies of the proposed method were achieved on an image data set collected from the bright-field microscopy screening for Mycobacterium tuberculosis in specimen of Ziehl–Neelsen-stained sputum smears.     

考虑圆形特征边缘模糊和偏心误差修正的高精度相机标定方法

针对立体视觉系统采用圆形特征点标定时存在的空间圆形投影边缘模糊和偏心现象问题,利用改进Zernike矩和偏心误差修正进行圆心的高精度定位,以此提高相机参数的标定精度。首先考虑了由于立体视觉成像系统的标定场景光照强度不均匀引起的圆形特征投影图像边缘模糊的问题,引入高斯误差函数对边缘过渡段的灰度分布进行描述,建立了高斯边缘模型,并基于该模型计算投影图像的Zernike矩,然后利用改进Zernike矩实现高精度的圆形特征投影边缘像素坐标定位。此外,分析了影响圆形特征中心投影点和拟合圆心间偏差大小的因素,基于该分析对迭代拟合圆心进行偏差补偿使之逼近真实的圆心投影,最后通过所提算法对99圆形标志点进行圆心坐标提取并用于相机参数的标定。仿真实验表明,文中算法对投影图像边缘定位的精度以及圆心拟合的精度均高于传统的算法;实测实验中,基于圆心高精度坐标得到的相机标定参数对标准杆进行三维重建,长度测量精度比传统算法提高了30%。     

基于双通道卷积神经网络的视频目标移除取证算法

针对现有数字视频目标移除取证算法的伪造帧识别准确率低的问题,本文提出了一种基于双通道卷积神经网络的视频目标移除取证算法。该算法利用双通道结构,分别提取视频绝对帧差图像的RGB特征和噪声特征,并利用双线性池化对二者进行特征融合,而后通过分类层输出视频帧的分类结果,从而有效地识别经过篡改的视频帧。其中,RGB通道能够发现绝对帧差图像中不自然的篡改边界和对比度,噪声通道能够发现原始区域和篡改区域之间噪声的不一致性。此外,算法在网络前端增加了预处理层来放大篡改视频帧的伪造痕迹。实验结果显示,所提算法有效地提高了伪造视频帧的识别准确率,且相对于传统的单通道网络结构,双通道特征融合的方式取得了更好的检测性能。      

Detection and localization of forgery using statistics of DCT and Fourier components

In this paper, we present a comprehensive approach for investigating JPEG compressed test images, suspected of being tampered either by splicing or copy-move forgery (cmf). In JPEG compression, the image plane is divided into non-overlapping blocks of size 8 × 8 pixels. A unified approach based on block-processing of JPEG image is proposed to identify whether the image is authentic/forged and subsequently localize the tampered region in forged images. In the initial step, doubly stochastic model (dsm) of block-wise quantized discrete cosine transform (DCT) coefficients is exploited to segregate authentic and forged JPEG images from a standard dataset (CASIA). The scheme is capable of identifying both the types of forged images, spliced as well as copy-moved. Once the presence of tampering is detected, the next step is to localize the forged region according to the type of forgery. In case of spliced JPEG images, the tampered region is localized using block-wise correlation maps of dequantized DCT coefficients and its recompressed version at different quality factors. The scheme is able to identify the spliced region in images tampered by pasting uncompressed or JPEG image patch on a JPEG image or vice versa, with all possible combinations of quality factors. Alternatively, in the case of copy-move forgery, the duplication regions are identified using highly localized phase congruency features of each block. Experimental results are presented to consolidate the theoretical concept of the proposed technique and the performance is compared with the already existing state of art methods.     

Target recognition in SAR images using radial Chebyshev moments

In this study, a new algorithm for classification of ground vehicles from standard synthetic aperture radar (SAR) images is proposed. Radial Chebyshev moment (RCM) is a discrete orthogonal moment that has distinctive advantages over other moments for feature extraction. Unlike invariant moments, its orthogonal basis leads to having minimum information redundancy, and its discrete characteristics explore some benefits over Zernike moments (ZM) due to having no numerical errors and no computational complexity owing to normalization. In this context, we propose to use RCM as the feature extraction mechanism on the segmented image and to compare results of the fused images with both Zernike and radial Chebyshev moments. Firstly, by applying different threshold target and shadow parts of each SAR images are extracted separately. Then, segmented images are fused based on the combination of the extracted segmented region, segmented boundary and segmented texture. Experimental results will verify that accuracy of RCM, which improves significantly over the ZM. Ten percent improvement in the accuracy is obtained by using RCM and fusion of segmented target and shadow parts. Furthermore, feature fusion improves the total accuracy of the classification as high as 6%.     

应用支持向量机分类的多角度目标识别技术

综合应用图像的不变矩特征和支持向量机分类方法,提出了一种对于红外图像中多角度目标的识别方法。首先通过目标分割算法求得红外图像中目标的轮廓图像,然后从轮廓图像的Hu矩、Zernike矩和Fourier-Mellin矩中选取适当阶次的矩特征组成目标在特定视角范围内的不变性特征向量;对目标的视角范围进行适当划分以解决多角度引起的目标样本多样性,并在每个划分的视角范围内分别应用支持向量机的方法进行多目标分类。测试结果表明,本文提出的方法较好地实现了红外图像中多角度目标的识别问题,是一种有效的自动目标识别算法。     

Robust image watermarking using local Zernike moments

In this work, we propose a robust image watermarking algorithm using local Zernike moments, which are computed over circular patches around feature points. The proposed algorithm locally computes Zernike moments and modifies them to embed watermarks, achieving robustness against cropping and local geometric attacks. Moreover, to deal with scaling attacks, the proposed algorithm extracts salient region parameters, which consist of an invariant centroid and a salient scale, and transmits them to the decoder. The parameters are used at the decoder to normalize a suspect image and detect watermarks. Extensive simulation results show that the proposed algorithm detects watermarks with low error rates, even if watermarked images are distorted by various geometric attacks as well as signal processing attacks.     

高速的9×9尺寸模板Zernike矩边缘算子

根据Zernike矩基本原理,提出了基于9×9模板尺寸的Zernike矩边缘算子。首先推导Zernike矩模板计算过程,然后计算出2个9×9模板,并选取二值图像和经过5次平滑的模糊图像为对象对新算子进行了测试。结果表明:对于理想的二值图像,新算子具有较强的边缘细化能力,且边缘检测精度也较高;而对于模糊图像,虽然新算子的边缘检测效果也较好,但应注意边缘图像表征与亚像素坐标数据间存在不一致的现象。最后,采用先粗后精定位思想,使新算子的运算速度大幅提高,耗时仅为0.19s。     

Design of complex adaptive multiresolution directional filter bank and application to pansharpening

This paper proposes a new 2-D transform design, namely complex adaptive multiresolution directional filter bank, to represent the spatial orientation features of an input image adaptively. The proposed design is completely shift invariant and represents the input image by one low-pass and multiscale N directional band-pass subbands. Here, N represents estimated number of dominant directions present in the input image. Our design consists of two main filter bank stages. A fix partitioned complex-valued directional filter bank (CDFB) is at the core of the design followed by a novel partition filter bank stage. Fine partitioning of the CDFB subbands is used to get the adaptive nature of the proposed transform. The partitioning decision is made based on the directional significance of range of CDFB subband angle selectivity in the input image. Partition filter bank stage which nonuniformly partitions the CDFB subbands provides total N dominant direction selective subbands. Local orientation map of the input image is used to determine the dominant directions and hence N. For better sparsity properties, we design the multiresolution stage with filters having high vanishing moments and better frequency selectivity. Applicability of the proposed adaptive design is shown for pansharpening of multispectral images. Our proposed pansharpening approach is evaluated on images captured using QuickBird and IKONOS-2 satellites. Results obtained using the proposed approach on these datasets show considerable improvements in qualitative as well as quantitative evaluations when compared to state-of-the-art pansharpening approaches including transform-based methods.     

Fusion of block and keypoints based approaches for effective copy-move image forgery detection

Keypoint-based and block-based methods are two main categories of techniques for detecting copy-move forged images, one of the most common digital image forgery schemes. In general, block-based methods suffer from high computational cost due to the large number of image blocks used and fail to handle geometric transformations. On the contrary, keypoint-based approaches can overcome these two drawbacks yet are found difficult to deal with smooth regions. As a result, fusion of these two approaches is proposed for effective copy-move forgery detection. First, our scheme adaptively determines an appropriate initial size of regions to segment the image into non-overlapped regions. Feature points are extracted as keypoints using the scale invariant feature transform (SIFT) from the image. The ratio between the number of keypoints and the total number of pixels in that region is used to classify the region into smooth or non-smooth (keypoints) regions. Accordingly, block based approach using Zernike moments and keypoint based approach using SIFT along with filtering and post-processing are respectively applied to these two kinds of regions for effective forgery detection. Experimental results show that the proposed fusion scheme outperforms the keypoint-based method in reliability of detection and the block-based method in efficiency.     

基于时频图像Zernike 矩特征的欺骗干扰识别

为了有效应对跟踪雷达的三种常见欺骗干扰,提出了一种基于平滑伪魏格纳-维尔分布时频图像的Zernike 矩特征的干扰识别方法。该方法对三种干扰下的雷达接收信号进行时频分析,运用数字图像处理技术对时频图像进行一系列的预处理后,通过Zernike 矩特征提取图像的细节特征组成特征向量进行分类识别。仿真实验证明:该方法有较高的识别率,特别是该方法受信噪比影响较小,能够有效降低噪声对干扰识别的影响,说明了采用图像识别方法对雷达欺骗干扰信号进行分类识别的可行性。通过与其他文献方法的比较,证明了该方法的优越性。     

基于组合矩的激光成像雷达目标识别算法

随着激光技术的发展,激光成像雷达在现代战争复杂战场环境中逐渐获得了广泛的应用,目前激光成像雷达自动目标识别技术已成为国内外研究的热点问题。提出了基于组合矩的激光成像雷达目标识别算法,从激光成像雷达目标的距离像中提取低阶的Zernike矩、Hu矩和中心矩构成组合矩特征,该特征对距离像噪声不敏感,应用径向基函数(RBF)神经网络对三种地面目标进行分类识别。实验结果表明,该算法与应用Zernike矩和Hu矩特征进行分类识别相比,对三种激光成像雷达地面目标的平均识别率在高载噪比(20dB)下分别提高了1.0%和3.7%;在低载噪比(10dB)下分别提高了11.8%和42.5%;当载噪比高于17dB时,该算法的平均识别率达到100%。因此该算法取得了比较好的识别效果。     

基于Zernike矩和熵的图像感知哈希算法

提出了基于Zernike矩和熵特征的数字图像感知哈希算法。算法利用Zernike矩计算参考方向,以计算等面积环块和等角度扇形块内的熵作为感知特征,并通过量化处理构造哈希序列。算法利用哈希码之间的欧氏距离作为图像内容相似性的判定依据。实验结果表明,该算法对加性噪声、JEPG压缩、几何变换等操作具有较好的鲁棒性,且对于内容不同的图像有较好的区分度。     

基于Harris特征点和伪Zernike矩的鲁棒水印算法

抵抗几何攻击的鲁棒性是目前数字水印技术中的热点亦是难点。文中以Harris检测算子和伪Zernike矩为理论基础,提出一种有效抵抗几何攻击的鲁棒水印算法。通过提取归一化图像的Harris特征点,选取部分稳定特征点确定圆形区域并计算伪Zernike矩,并量化调制伪Zernike矩的幅值来嵌入水印信息。实验结果表明,本算法对于抗几何攻击具有较好的鲁棒性,同时也能抵抗常规的信号处理攻击。     

基于深度特征提取和图神经网络匹配的图像复制粘贴篡改检测

针对图像中特征提取不均匀、单尺度超像素划分对伪造定位结果影响较大的问题,提出一种基于深度特征提取和图神经网络(graph neural network,GNN) 匹配的图像复制粘贴篡改检测(cope-move forgery detection,CMFD) 算法。首先将图像进行多尺度超像素分割并提取深度特征,为保证特征点数目充足,以超像素为单位计算特征点分布的均匀度,自适应降低特征提取阈值;随后引入新的基于注意力机制的GNN特征匹配器,进行超像素间的迭代匹配,且用随机采样一致性(random sample consensus,RANSAC) 算法消除误匹配;最后将多尺度匹配结果进行融合,精确定位篡改区域。实验表明,所提算法具有良好的性能,也证明了GNN在图像篡改检测领域的可用性。     

Exact Zernike and pseudo-Zernike moments image reconstruction based on circular overlapping blocks and Chamfer distance

This study aims to explore a novel approach to reconstruct multi-gray-level images based on circular blocks reconstruction method using two exact and fast moments: Zernike (CBR-EZM) and pseudo-Zernike (CBR-EPZM): An image is first divided into a set of sub-images which are then reconstructed independently. We also introduced Chamfer distance (CD) to capitalize on the use of discrete distance instead of Euclidean one. The combination of our methods and CD leads to CBR-EZM-CD and CBR-EPZM-CD methods. Obviously, image partitioning offers significant advantages, but an undesirable circular blocking effect can occur. To mitigate this effect, we have implemented overlapping feature to our new methods leading to OCBR-EZM-CD and OCBR-EPZM-CD, by exploiting neighborhood information of the circular blocks. The main motivation of this novel approach is to explore new applications of Zernike and pseudo-Zernike moments. One of the fields is feature extraction for pattern recognition: Zernike and pseudo-Zernike moments are well known to capture only the global features, but thanks to the circular block reconstruction, we can now use those moments to extract also local features.