基于横向色差一致性的图像真伪鉴别

Identifying Image Authenticity Through Lateral Chromatic Aberration Consistency

随着图像处理技术的不断成熟,涌现出越来越高级的图像处理工具,使得人们对篡改的图像很难用肉眼鉴别。由于相机在成像过程中受到光学系统的制约导致图像存在色像差,可以利用这一特性鉴别图像的真伪。提出一种基于横向色差一致性的图像篡改检测方法,该方法根据通道间的块匹配估计横向色差的分布,基于亮度极值将图像划分为不同的同心圆区域,在不同区域内度量估计的横向色差分布与理论分布的差异,结合横向色差的强度信息采用加权的方法确定图像的整体差异,通过阈值判断得到图像的篡改检测结果。实验结果表明,该方法提高了检测的准确率,具有良好的鲁棒性,从而可以有效地鉴别图像真伪。

With the continuous maturity of image processing technology,more and more advanced image processing tools have emerged,making it difficult for people to visually detect forgery images.Since the camera causes the emergence of chromatic aberration during the imaging process,this can be used to identify the authenticity of the image.This paper proposes a detection method based on lateral chromatic aberration consistency.This method uses the block matching across channels to estimate lateral chromatic aberration distribution.The image is divided into different concentric circles according to the extreme values of the intensity.The difference between the lateral chromatic aberration estimates and the theoretical distribution is measured in different regions.Combined with the intensity of lateral chromatic aberration,the overall difference is determined by the weighting method.The forgery detection result is obtained using the threshold value.The experimence results show that this method improves the accuracy of detection and has strong robustness,which can identify the authenticity of the image effectively.