摘要
针对篡改造假图像中常用的模糊操作,提出了基于边缘模糊检测的伪造图像盲鉴别算法。经过对模糊原理的研究发现,图像经过二次或者多次模糊后,模糊区域的像素差值小于正常区域。该算法首先对待检测图像利用一定的模糊函数进行全局模糊;然后对模糊前后的图像进行差值计算,得到差值图像;最后根据差值图像中像素值的异常区域进行篡改定位。实验结果表明,该算法对局部模糊和拼接合成图像边缘模糊两大类篡改都能进行有效鉴别,抗JPEG压缩的能力较强;相对于先前的一些算法,提高了检测率并且降低了算法复杂度。
To identify blur operation that is commonly used in the tampered and forged images, a blind identification algorithm based edged blur in a forged image was proposed. After a study on the principle of the fuzzy image, the author found that the pixcl difference value in the blurred region is less than the normal region after twice or multiple blur. Therefore, firstly, the algorithm in this paper did a global blur to the tested image by blur function; secondly, the difference value between tested image and blurred test image was calculated, and get the difference image; At last, the tampered abnormal area was located according to the difference value in the image. The experimental results show that the proposed algorithm can effectively identify local blur of image and edge blur of split-compose image, and the ability to resist JPEG compression is higher. Compared to the previous algorithms, the algorithm improves the detection rate and reduces the algorithm complexity.
出处
《计算机应用》
CSCD
北大核心
2013年第A02期159-161,共3页
journal of Computer Applications
基金
国家自然科学基金资助项目(60773098)
关键词
图像篡改
模糊
像素差值
篡改检测
定位
image tamper
blur
pixel difference
tamper detection
location
