基于模糊度和色调变化率的图像模糊取证

FORENSICS OF IMAGE BLUR BASED ON BLUR DEGREE AND HUE CHANGING RATE

提出一种数字图像模糊篡改的检测方法。由于模糊操作是一种常用的图像润饰篡改手段,所以检测人工模糊对图像的真实性和原始性具有一定的必要性。首先利用图像的模糊度,估计出图像中可能经过模糊篡改的区域,再根据模糊操作会破坏图像邻域的色彩一致性且被破坏后的色彩一致性无法被"再次"破坏的性质对估计出的模糊区域进行高斯低通滤波处理,然后根据滤波前后色调值的变化个数进行精确的检测,最后定位出人工模糊篡改区域。实验表明,该方法可以检测出图像中的人工模糊篡改区域,并能准确地将其进行定位。

In this paper we proposed a digital image blur tampering detection method. Since blur operation is a commonly used means in image retouching and tampering, so to detect artificial blurring on the reality and primitiveness of image has certain necessity. First, the method employs the blur degree of image to estimate the regions in image possibly tampered in blurring way, then according to the property of blur operation that it will destroy fhe colour consistency of image neighbourhood and the destroyed eolour consistency cannot be destroyed ＂once again＂, the method carries out Gaussian low-pass filtering process on the estimated blurting regions, and makes accurate detection according to the changed numbers of hue before and after filtering, and finally locates the artificially blur tampered regions. Experiment shows that the proposed method can detect the artificially blur tampered regions in image, and can accurately locate them.