标题:
一种基于频率的复制粘贴图像篡改检测A Detection Method for Copy-Move Image Forgery Using Frequency-Based Techniques
作者:
胡美栋, 王孝通, 徐冠雷, 于浩
关键字:
图像篡改, 基于频率, 复制粘贴篡改, 数字取证Image Forgery, Frequency-Based, Copy-Move Forgery, Digital Forensics
期刊名称:
《Journal of Image and Signal Processing》, Vol.6 No.1, 2016-12-15
摘要:
由于可访问性的图像编辑软件的普遍存在，数字图像正在逐渐失去它们的可信度。为了检验数字图像的真实性，图像篡改检测方法必不可少。最常见的图像篡改方法之一就是复制粘贴篡改。检测这种类型的篡改的主要原则是基于发现不同图像段之间的相似性。本文提出了一种基于频率的图像检测方法，用于检测图像中的复制粘贴篡改。先将图像分割成不同的块，然后提取出各个图像块的特征向量。通过匹配和滤波，两个相似的特征向量分别对应图像中两个相似的区域，即复制粘贴篡改区域。
Due to ubiquitous accessibility of image editing software, digital images are losing their credibility. For checking authenticity of digital images, forgery detection methods are required. One of the most frequent image forgeries is copy-move. The main principle of detecting such type of forgery is based on finding resemblance present in different segments of image. In this paper, a survey on frequency-based methods is presented for detecting copy-move forgery in images. Image is divided in blocks, and then feature vectors are extracted corresponding to different blocks of image. By matching and filtering, similar feature vectors are configured to find similar regions of image, and those are copy-move regions.