Forgery Detection Research Based On Digital Images And Videos

With the development of information technology, there are more and more typesof digital image editing software with perfect function. Ordinary users do not haveprofessional computer knowledge can also modify and forgery the images. Withwidely used of network application, spread the doctored photos have a certain impacton the government departments and the people’s life. Images forge detection researchwill be more important and a topic has good prospects.This paper analyzes several main methods of current digital image forgery.Combining the algorithm already exist and proposed by this paper to detect the forgeimage. We also propose a video capturing device identification approach by exploringthe H.264cues to determine whether a video has been modified from its originalrecording. In this paper, considering the diversity of digital image format, forgingmethods of concealment. Our system use geometric information, Google street viewand EXIF information to judge whether this image has been doctored. Our algorithmalso exploits image feature information, Such as SURF feature. When we detectedimage use Google street view, we first extract EXIF data. Then we can find the streetview location with GPS. Lastly, we can judge whether this image has been edited. Forvideo capture device identification we use H.264encode parameters cues.This paper mainly focus on media data forge detection, Including image forgedetection and video original source. Our method based on active analysis, will notaffect the image or video structure. For the fake pictures from network, our algorithmgives a good result.