CAJ | 학술논문

为在像素级上提高对图像拼接区域的定位精度，分析图像中自然边缘像素点和拼接边缘像素点与其邻域像素点之间存在的平滑过渡关系，提出一种有效的图像拼接盲取证定位算法。对图像中所有的边缘像素点提取平滑度相关性标量值，将其与判决阈值进行比较，得到拼接边缘，定位拼接区域。在标准拼接图像数据集中的实验结果表明，该方法能够有效提高图像拼接区域的定位精度。
위재상소급상제고대도상병접구역적정위정도，분석도상중자연변연상소점화병접변연상소점여기린역상소점지간존재적평활과도관계，제출일충유효적도상병접맹취증정위산법。대도상중소유적변연상소점제취평활도상관성표량치，장기여판결역치진행비교，득도병접변연，정위병접구역。재표준병접도상수거집중적실험결과표명，해방법능구유효제고도상병접구역적정위정도。
To improve accuracy of locating image splicing region in pixel level,a blind forensics approach of splicing location based on principle of color filter array interpolation in camera imaging was proposed.Smoothness of the neighborhood pixels of natural edge pixels and splicing edge pixels in the image introduced by color filter array interpolation was analyzed.Splicing edge was extracted after comparing smoothness correlation scalar value of all the edge pixels with the threshold obtained.The results of experiment on the splicing image data set show that the proposed method can effectively improve accuracy of locating the image splicing area.