Abstract: After learning the image matching algorithm by combining Scale Invarian Feature Transform(SIFT) and the rotation-invariant Local Binary Patterns(LBP), this paper proposes a detection algorithm of image region copy-move forgery based on LBP. All keypoints are extracted from the image by applying the SIFT algorithm, and each keypoint is described by the rotation-invariant LBP, which are computed from the image patch centered at the keypoint. It matches the keypoint by using the feature vectors of Euclidean distance. Experimental results show that this algorithm performs better than the classical detection algorithm based on Principal Component Analysis(PCA) in terms of robustness against rotate operation, intensity adjustment and efficiency.

Key words: digital image forensics, image region copy-move forgery, Scale Invarian Feature Transform(SIFT), Local Binary Patterns(LBP)

摘要： 研究尺度不变特征变换(SIFT)和旋转不变局部二值模式(LBP)相结合的特征匹配方法，提出一种基于LBP的图像区域复制-粘贴篡改检测算法。利用SIFT关键点检测方法检测图像中的所有关键点，计算以关键点为中心的周围图像区域的LBP特征，并将其作为关键点的特征描述，采用特征向量的欧式距离进行关键点匹配。实验结果表明，该算法在抗旋转、亮度变化处理和效率方面均优于基于主成分分析的检测算法。?

关键词: 数字图像取证, 图像区域复制-粘贴篡改, 尺度不变特征变换, 局部二值模式

CLC Number: 

• TN911.73