摘要
基于内容的图像拷贝检测已经成为图像版权保护的重要手段之一。提出一种基于图像边缘直方图的拷贝检测方法:采用小波模极大值算法提取源图像和待认证图像边缘,然后结合边缘点梯度主方向重新计算其区域点的梯度方向并生成直方图,最后根据源图像与待认证图像所得边缘直方图间的相似性检测待认证图像是否为源图像的拷贝。实验表明:这种图像拷贝检测方法可有效检测出拷贝图像,并可容忍由缩放、旋转、噪声等操作引起的失真。
Content-based image copy detection has become one of the important issues of image copyright protection.In this paper, an edge histogram-based image copy detection algorithm is presented.At first, wavelet modulus maxima algorithm is used to extract the edge feather of the source image to identify the candidate images.Then the gradient direction is re-calcu- lated by combining with the main direction of the edge point gradient and computes the corresponding edge histogram.Final- ly it is judged that whether the candidate images are the copies of the source image by computing the image edge histo-gram similarities between the source image and the candidate images.Experimental results verify that the proposed method can effectively detect copied images, and tolerant the scaling,rotating,noise-induced distortion operation.
出处
《计算机工程与应用》
CSCD
北大核心
2010年第31期184-187,共4页
Computer Engineering and Applications
基金
国家科技支撑计划(No.2007BAH14B05)
湖南省教育厅优秀青年项目(No.06B095)
关键词
图像拷贝检测
边缘检测
主方向
image copy detection
edge detection
main direction
