图像拼接篡改的自动色温距离分类检验方法

拼接篡改是一类常见的图像伪造手段,现有取证方法难以实现图像中拼接篡改区域的自动检测与精确定位,导致拼接篡改伪造图像的取证长期依赖人工经验.基于图像中原始区域与拼接篡改区域所反映的光源色温的差异性,提出一种自动色温距离阈值分类的图像拼接篡改检测与定位方法.首先,变换待检验图像至YCbCr色彩空间,并按照Grid-based方式结构化分解为大小的子图像块;然后,利用自动白平衡（Automatic white balance,AWB）中的白点检测原理对每一个子图像块进行色温估计,计算子图像块与参考区域之间的色温距离;最后,采用最大类间方差法自适应地求取色温距离分类的最佳阈值,对子图像块进行分类标注,实现了图像拼接篡改区域的自动检测与精确定位.实验表明,该方法能够实现图像拼接篡改区域的自动检测与定位,具有较高的量化检测精度.     

基于光源颜色和噪声的图像拼接篡改检测方法

数字图像拼接篡改是一种常见的图像伪造手段。在图像取证中,检测拼接伪造仍然是一项艰巨的任务。现有的拼接伪造检测方法多根据一种图像特性的不一致检测并定位篡改区域,而实际拼接篡改伪造往往会造成多种图像特性的改变。针对现有单一特征提取不能充分反映拼接图像特性导致检测精确率低的问题,提出一种通过提取光源颜色和噪声的混合特征来定位并显示拼接区域的高效图像拼接检测方法。实验结果表明,混合特征提取方法较单一特征提取方法能取得更高的检测精确率。     

USM增强的边缘羽化拼接图像检测方法

图像拼接过程中,经过羽化模糊处理的图像边缘特征是不同于自然图像边缘特征的。因此,可利用图像边缘羽化特征来进行图像拼接区域的检测。然而,当图像中某些边缘与羽化边缘特征相似时,直接提取边缘羽化特征的方法会导致误判。如果能够扩大拼接区域的边缘与自然图像边缘之间的差异,则可以更准确地检测出图像中的篡改区域。因此,提出了一种基于USM增强的边缘羽化拼接图像检测方法。首先,对待检测的拼接图像进行UMS增强,扩大拼接图像中羽化边缘与自然边缘特征的差异性。然后,计算增强后图像边缘像素的羽化半径。最后,寻找半径相似的边缘像素,定位图像中的拼接篡改区域。实验结果表明,在拼接图像中存在与羽化边缘相似的边缘时,该方法能更准确地检测出拼接区域。     

面向图像篡改取证的多特征融合U形深度网络

随着图像篡改工具的智能化发展,图像篡改不再局限于拼接、移除等某一具体的类型,往往包含多种篡改类型及其组合操作,使得图像篡改取证工作更具挑战性。提出一种端到端的多特征融合U形深度网络,利用编解码网络提取篡改区域与真实区域之间的对比度差异、边缘差异等篡改痕迹,并使用富隐写模型卷积层获取伪造图像的噪声分布不规律信息,从而在无预处理的情况下实现可疑区域的检测并分割出高置信度的篡改区域。在此基础上,使用特征提取模块获取融合的篡改特征,在融合定位模块中利用分级监督策略融合不同分辨率提取的篡改特征,以准确定位篡改区域,实现篡改区域检测与像素级的分割。实验结果表明,基于所提网络的图像篡改取证方法在NIST16和CASIA数据库上的F1值分别为0.841和0.605,与基于MFCN、RGB-N、MANTRA-net等网络的图像篡改取证方法相比,有较优的检测性能和较高的实时性,且对JPEG压缩、缩放等处理具有更强的鲁棒性。     

一种改进的双流Faster R-CNN图像篡改识别模型

针对目前大多数图像篡改算法只能针对一类图像篡改进行检测，以及双流Faster R-CNN算法提取的RGB流和噪声流特征对多种图像篡改检测精度不高的问题，提出一种通用的基于改进的双流Faster R-CNN图像篡改识别算法。提取图像的YCrCb颜色空间，代替之前的RGB颜色空间，以更好找出篡改的痕迹；对提取噪声特征的三个隐写分析丰富模型(SRM)滤波器进行旋转变换，以更好区别真实区域和篡改区域的噪声不一致，从而提高对篡改图像的识别精度；通过双线性池化，输入网络训练和分类，完成对图像篡改的检测与篡改区域定位。为验证算法的性能，在CASIA和NISIT16两个数据集上进行了实验。结果表明，与双流Faster R-CNN算法相比，提出的图像篡改识别算法在拼接检测、复制移动检测和移除检测上平均精度(AP)分别提升0.9百分点、1.5百分点和2.6百分点。     

拼接篡改伪造图像的色彩偏移量不一致取证方法

拼接篡改是一种常见的伪造图像方法,根据拼接篡改伪造图像中拼接区域与原始区域之间存在的色彩偏移量的差异,提出一种基于偏色估计的拼接篡改伪造图像自动检测方法.首先将图像分为n×n大小的图像子块,利用改进的平均色差计算方法估计每一个子块的色彩偏移量;然后将分块之后的待检验图像的上、左、右3个径向方向上的尺度边缘子块的集合设定为参考区域;计算每一个子块与参考区域之间的偏色距离,最后与设定的偏色距离阈值进行比较后定位图像中的拼接区域,从而揭示拼接篡改图像中存在的色彩偏移量不一致现象.实验结果表明,该方法能够自动检测拼接篡改图像中的色彩偏移量不一致并定位拼接篡改区域,为拼接篡改伪造图像的取证提供了一类科学量化的检验依据.     

采用圆谐-傅里叶矩的图像区域复制粘贴篡改检测

现有检测方法大多对图像区域复制粘贴篡改的后处理操作鲁棒性不高.针对这种篡改技术,提出一种新的基于圆谐-傅里叶矩的区域篡改检测算法.首先将图像分为重叠的小块;然后提取每个图像块的圆谐-傅里叶矩作为特征向量并对其进行排序;最后根据阈值确定相似块,利用位移矢量阈值去除错误相似块以定位篡改区域.实验结果表明,该算法能有效抵抗噪声、高斯模糊、旋转等图像后处理操作,且与基于HU矩的方法相比有更好的检测结果.     

基于量化噪声的JPEG图像篡改检测

JPEG图像篡改引入的双重压缩会导致篡改区域的原始压缩特性发生改变,因此可以利用篡改区域压缩特性的不一致性来检测图像的篡改。利用该原理,提出了一种基于量化噪声的JPEG图像篡改检测算法。算法对待检测图像进行分块,计算每块的量化噪声,求取图像块的量化噪声服从均匀分布和高斯分布的概率,从而检测出篡改过的双重压缩区域。实验结果表明：该算法能有效检测双重压缩的JPEG图像篡改,并能定位出篡改区域。     

基于背景噪声的图像盲篡改检测

数字图像在成像过程中会产生特定的背景噪声,如果两幅不同噪声的图像拼接在一起,篡改区域和其他区的噪声会有差异。提出一种基于偏度统计特性的背景噪声估计算法,其通过对图像分块计算每块的噪声标准差,从而检测出噪声异常部分以达到篡改检测的目的。算法利用DCT变换去除原图细节部分,利用偏度统计特性估计噪声,利用条件最小值法求出噪声的标准差。算法改进了迭代求条件最小值法,利用微分方法求取最小值,避免了初始值设定问题,提高了算法的准确率。实验结果表明,提出的噪声估计算法正确率高,且对拼接篡改图像篡改检测有明显效果。     

基于Transformer的多任务图像拼接篡改检测算法

现有基于深度学习的图像拼接篡改检测方法大多依赖卷积操作的局部计算过程,感受野有限。此外,现有方法大多仅将篡改区域定位用于指导检测模型训练,难以学习更加丰富的篡改痕迹特征。针对上述局限性,提出了基于Transformer的多任务图像拼接篡改检测网络(Multitask Transformer-based Network, MT-Net),利用Transformer中的自注意力机制在特征提取过程获取图像像素之间的相关性,自适应地为各像素提供不同的关注度,提升检测网络对篡改痕迹的表征能力。此外,MT-Net同时考虑多个子任务从局部细化和整体感知两个方面共同引导网络学习,包括篡改区域定位、篡改边缘定位和篡改比例预测,并根据子任务特点设计了对应的损失函数来指导网络进行优化。实验结果表明,相比现有算法,所提算法在CASIA V2.0,Columbia和IDM2020这3个公开数据集上均取得了更好的检测准确性,F1值分别达到了0.808,0.913和0.675。可视化检测结果图表明,所提算法在定位拼接篡改区域时也有较好的表现。     

Image splicing localization using noise distribution characteristic

Image splicing/compositing is common content tampering operation. In this work, we devote to improve the detection accuracy of the splicing/compositing attack for image, and propose an effective image splicing localization method based on the noise distribution characteristic in image. Firstly, the test image is divided into non-overlapping blocks by using an improved simple linear iterative clustering (SLIC) algorithm. Then block-wise local noise level estimation and noise distribution characteristic estimation are performed to generate distinguishing features. Utilizing the fact that image regions from different sources tend to have larger inter-class difference, the fuzzy c-means clustering is used to identify spliced regions. Compared to existing noise-based image splicing detection methods, experimental results on different datasets have shown that the proposed method has superior performance, especially when the noise difference between the spliced region and the original region is small. Moreover, the proposed method is robust for content-preserving manipulations.

     

Exposing Region Splicing Forgeries with Blind Local Noise Estimation

Region splicing is a simple and common digital image tampering operation, where a chosen region from one image is composited into another image with the aim to modify the original image’s content. In this paper, we describe an effective method to expose region splicing by revealing inconsistencies in local noise levels, based on the fact that images of different origins may have different noise characteristics introduced by the sensors or post-processing steps. The basis of our region splicing detection method is a new blind noise estimation algorithm, which exploits a particular regular property of the kurtosis of nature images in band-pass domains and the relationship between noise characteristics and kurtosis. The estimation of noise statistics is formulated as an optimization problem with closed-form solution, and is further extended to an efficient estimation method of local noise statistics. We demonstrate the efficacy of our blind global and local noise estimation methods on natural images, and evaluate the performances and robustness of the region splicing detection method on forged images.     

多工件拼接焊缝面结构光滑度的视觉检测技术研究

由于多工件拼接焊缝面结构光滑度检测过程中,受其拼接面复杂且图像存在噪声,导致细节特征不明显,从而使检测精度、效率降低。为此,提出多工件拼接焊缝面结构光滑度的视觉检测技术。采用字典学习方法消除焊缝面图像中存在的噪声；将其输入MRFENet网络,提取图像特征,实现焊缝面图像的增强处理；采用增量二维主成分分析提取焊缝面结构的光滑度特征,实现光滑度检测。实验结果表明,所提方法图像处理效果好、检测精度高、检测效率高。     

Detecting Image Splicing Based on Noise Level Inconsistency

In a spliced image, areas from different origins contain different noise features, which may be exploited as evidence for forgery detection. In this paper, we propose a noise level evaluation method for digital photos, and use the method to detect image splicing. Unlike most noise-based forensic techniques in which an AWGN model is assumed, the noise distribution used in the present work is intensity-dependent. This model can be described with a noise level function (NLF) that better fits the actual noise characteristics. NLF reveals variation in the standard deviation of noise with respect to image intensity. In contrast to denoising problems, noise in forensic applications is generally weak and content-related, and estimation of noise characteristics must be done in small areas. By exploring the relationship between NLF and the camera response function (CRF), we fit the NLF curve under the CRF constraints. We then formulate a Bayesian maximum a posteriori (MAP) framework to optimize the NLF estimation, and develop a method for image splicing detection according to noise level inconsistency in image blocks taking from different origins. Experimental results are presented to show effectiveness of the proposed method.     

Image splicing localization using PCA-based noise level estimation

Image splicing is one of the most common image tampering operations, where the content of the tampered image usually significantly differs from that of the original one. As a consequence, forensic methods aiming to locate the spliced areas are of great realistic significance. Among these methods, the noise based ones, which utilize the fact that images from different sources tend to have various noise levels, have drawn much attention due to their convenience to implement and the relaxation of some operation specific assumptions. However, the performances of the existing noise based image splicing localization methods are unsatisfactory when the noise difference between the original and spliced regions is relatively small. In this paper, through incorporation of a recent developed noise level estimation algorithm, we propose an effective image splicing localization method. The proposed method performs blockwise noise level estimation of a test image with principal component analysis (PCA)-based algorithm, and segments the tampered region from the original region by k-means clustering. The experimental results demonstrate the superiority of the proposed method over several state-of-the-art methods, especially for practical image splicing, where the noise difference between the original and spliced regions is typically small.     

基于高频小波子带马尔可夫特征的图像拼接检测

拼接是图像篡改过程中最普遍使用的操作,通过检测拼接可以有效鉴别图像是否经过人为修改。针对拼接操作提出了一种盲检测方法：首先对图像进行小波变换,在比较分析不同小波子带对图像拼接检测的作用后,选取高频子带作为图像变换域信息;接着对小波子带进行差分操作,并将系数取整阈值化后作为马尔可夫状态;最后计算状态间的转移概率作为拼接特征,利用支持向量机（SVM）进行分类。在哥伦比亚图像拼接评测彩色库和灰度库上分别进行实验,证实了选取小波高频子带提取拼接特征的有效性。通过与其他特征提取方法对比,所提出特征在两个评测库上都表现出了更好的检测效果,尤其在彩色评测库上取得了94.6%的识别率。     

Quantization-based Markov feature extraction method for image splicing detection

In this paper, we propose an efficient Markov feature extraction method for image splicing detection using discrete cosine transform coefficient quantization. The quantization operation reduces the information loss caused by the coefficient thresholding used to restrict the number of Markov features. The splicing detection performance is improved because the quantization method enlarges the discrimination of the probability distributions between the authentic and the spliced images. In this paper, we present two Markov feature selection algorithms. After quantization operation, we choose the sum of three directional Markov transition probability values at the corresponding position in the probability matrix as a first feature vector. For the second feature vector, the maximum value among the three directional difference values of the three color channels is used. A fixed number of features, regardless of the color channels and test datasets, are used in the proposed algorithm. Through experimental simulations, we demonstrate that the proposed method achieves high performance in splicing detection. The average detection accuracy is over than 97% on three well-known splicing detection image datasets without the use of additional feature reduction algorithms. Furthermore, we achieve reasonable forgery detection performance for more modern and realistic dataset.     

基于自适应四元数奇异值分解的图像拼接检测

图像拼接将来源不同的图像合并成一幅图,由此引起图像中光照方向、噪声等特性出现不一致的情况。目前多数方法根据拼接图像中噪声的不一致性来检测伪造区域,但是普遍对不同大小图像块的噪声估计准确性不高,导致真阳性率较低,且当噪声差异较小时会检测失败。针对该问题,提出一种基于自适应四元数奇异值分解（QSVD）的噪声估计方法。对图像进行超像素分割,利用自适应QSVD估计超像素的噪声,结合图像亮度并利用多项式拟合建立图像噪声-亮度函数,得到各超像素到该函数曲线的最小距离测度。为提高检测精确率,利用色温估计算法提取超像素的色温特征,将距离测度与色温特征相融合作为最终的特征向量,利用FCM模糊聚类定位拼接区域。在Columbia IPDED拼接图像数据集上进行实验,结果表明,该方法在未经后处理图像集上的检测TPR值较对比方法至少提升8.21个百分点,且对高斯模糊、JPEG压缩和伽马校正表现出较好的鲁棒性。