数学形态学在数字伪造模糊检测中的应用

Blur detection of digital forgery using mathematical morphology

针对图像篡改中常用的模糊操作,研究使用平滑滤波保存边和数学形态法的新型模糊检测方案,充分利用2种方法的锐化功能和去噪功能,无需水印技术等任何嵌入信息就能指出可能的篡改并定位篡改位置,该方法不仅可以判断某个图像是否被虚化过,还能检测该图像的虚化程度。实验结果证实了该方案的有效性。

Aiming at the fuzzy operation commonly used in image tampering, a new fuzzy detection scheme using smooth filter preserving edges and mathematical morphological methods was studied. By using the sharpening function and denoising function of both methods, tampering and location of tampering position can be pointed out without em- bedding information such as watermarking technique. The method can not only determine whether an image has been virtualized, but also detect the extent of the image blur. Experimental results confirm the effectiveness of the scheme.