作　　者： (王志红）;

机构地区： 山东英才学院,济南250104

出　　处： 《激光杂志》 2017年第8期118-121,共4页

摘　　要： 当前图像篡改检测方法多基于块效应、方法复杂,对于纹理丰富、色彩多变的篡改图像检测率较低、而且容易造成对图像篡改的误判。提出改进图像篡改检测方法。利用激光脉冲峰对图像进行阈值分割处理,然后提取待检测篡改图像的特征点,构建图像区域篡改模型,实现图像特征向量和阈值合理性分析和分解模式匹配,实现对篡改区域的判定和检测。实验结果表明,本文方法能够精确识别出邻近像素的特征差异点,有效提高图像篡改的检测率,降低了误判率。 Current method for image tamper detection is complex and based on blocking effect. And the detection rate is low for rich texture image with many clolors. So it is easy to cause the misjudgments. Therefore,an improved image tamper detection method is proposed. Laser pulse peak is used for threshold segmentation of image. And then extract the feature points of tampering images which are ready to be detected,and construct image tampered region model to realize the rational analysis of image feature vector and threshold,and the match of decomposition model. Finally implement the identification and detection of tampering area. The experimental results show that the proposed method can accurately identify the characteristic differences between adjacent pixels,effectively improve image tamper detection rate and reduce false rate.

关 键 词： 脉冲技术 图像篡改 检测 误判率