Research On Image Splicing Localization Algorithm Based On Super Pixel Segmentation

With the rapid development of computer technology and multimedia technology,photography has become a part of people's daily life.The growing digital technology makes it easier and easier to take pictures.And the wide application of various digital image modification software makes people more convenient to amend the image.As a result,the number of tampered images in life is increasing in blowout.Image splicing,as an important means of image tampering,is widely used in daily life.The splicing image has brought a lot of joy to people,but it also brings some negative effects.The existence of a large number of splicing images makes people more suspicious of the authenticity of the image information,more and more problems are highlighted.In some cases,the authenticity of images is very important.Therefore,people urgently need an effective way to determine whether the image has been tampered with,so the image tampering blind detection technology is born.However,some images that are badly tampered need people to grasp the exact area of being tampered with.Therefore,many scholars have done in-depth research on image splicing localization algorithm in image tampering blind detection technology.In view of the problem of low accuracy and high miscarriage rate,a lot of research has been done in this paper.First of all,through the background and significance of image forensics technology and image matching detection algorithm,this paper affirms the importance of the research;secondly,detailed analysis of research status at home and abroad about image splicing localization algorithm,and classification are introduced;finally,based on the thorough study of the model,put forward two kinds of detection algorithm for image splicing localization algorithm based on super-pixel.The research work of this subject is as follows:1.In order to solve the problem of low accuracy and high computational complexity,a new image splicing localization algorithm is proposed in this paper.First,the SLIC super pixel segmentation algorithm is used to segment the detection image.Secondly,the FAST noise estimation algorithm is used to calculate the noise estimation of each super pixel block.Finally,the noise value sequence is clustered and processed to determine the super pixel block in the background area or the splicing area,so as to locate the splicing area.The algorithm is experimentation on the color splicing image library of Columbia University,and compared with the existing image splicing localization algorithm based on block segmentation and pixel based.The experimental results show that the algorithm can effectively improve the efficiency of location detection on the premise of ensuring the accuracy of image stitching location detection.2.According to the leakage detection and error detection problems of splicing localization algorithm based on super pixel segmentation and noise estimation,based on the above methods,we integrate the idea of RC algorithm and consider the influence of spatial distance between pixels and color similarity,and further study the algorithm of image splicing localization.First of all,still use SLIC super pixel segmentation algorithm to segment image block and estimate noise;secondly,calculate the pixel block size,the color similarity and spatial distance between the pixel block;thirdly,weighting calculate the pixel block size,the spatial distance between pixel blocks,color similarity and noise feature;finally,according to the weights of all pixel blocks of two value classification,thus,locate the splicing area in the image.The experimental results show that this method can further improve the accuracy of image splicing localization.