局部特征在数字图像同源同源篡改取证中的应用研究分析研究.docxVIP

PAGE PAGE 1 破损、锁屏、停用手机取证方法研究 PAGE PAGE 2 局部特征在数字图像同源同源篡改取证中的应用研究 摘要 随着计算机和网络技术的不断发展，数字图像资源的使用率居高不下。在图像媒体得到广泛应用的同时，应各类图像使用的要求，图像处理技术飞速发展，出现了很多图像编辑软件，同源篡改图像变得相当容易，导致图像的安全性受到严重挑战。因此，用于图像信息安全的相关认证技术应用而生，并且迅速发展。 克隆操作和拼接操作可以看做在同一副图像或不同图像中生成了原始图像中部分图像的拷贝，可以考虑用局部特征拷贝检测的思想解决这两张操作的检测问题。因此，本文提出了基于局部特征拷贝检测的盲取证算法。一方面，通过特征融合提取精准特征，保证检测的准确性;另一方面，通过基于机器学习的感知哈希将特征映射为紧凑的哈希码，提高检测的效率。 但展望海量数字图像在互联网高速传播的未来，仍需继续完善数字图像盲取证技术框架、探索数字图像真实性相关的图像语义、研究利用互联网用户反馈信息帮助数字图像取证的方法等来提高面向真实性鉴别的数字图像盲取证技术的性能。 关键字：局部特征；数字图像；同源同源篡改；取证  ABSTRACT With the continuous development of computer and network technology, the use rate of digital image resources is high. With the wide application of image media, various image requests and image processing technologies have been developing rapidly. Many image editing softwares have appeared. The image of homologous tampering has become quite easy, which leads to serious challenges to image security. Therefore, the relevant authentication technology for image information security has been developed and developed rapidly. Cloning operation and stitching operation can be regarded as copies of original images in the same image or different images. We can consider the idea of local feature copy detection to solve the detection problems of these two operations. Therefore, a blind forensics algorithm based on local feature copy detection is proposed in this paper. On the one hand, accurate features are extracted through feature fusion to ensure the accuracy of detection. On the other hand, the perception of Hashi based on machine learning maps the feature to compact Hashi code, which improves the efficiency of detection. But the prospect of massive digital image on the Internet propagation speed of the future, still need to continue to improve the digital image forensics technology framework, explore the authenticity of digital image, research related to the use of the Internet user performance feedback method of digital image forensics help to improve the blind