基于LBP的同图复制粘贴篡改图像的检验研究. (Chinese)

With the rapid development of science and technology and the Internet, digital images are playing a more indispensable role in our life. Digital images are increasingly inseparable from social media, forensics, journalism, scientific research, medicine, and all sorts of industries. However, due to the high editability of digital image, anyone can modify the content of digital image arbitrarily through the image editing program, which makes the authenticity and objectivity of the image become unknown. Therefore, how to identify the authenticity of the image has become a hot topic among experts and scholars. In the recognition of image authenticity, the tampering form of modifying the image content by copying and pasting the same image is very common in many cases of image tampering. Aiming at such problems, this paper attempts to identify the copy-pasted tampering image of the same image based on LBP algorithm. Firstly, the corner points in the tampered image were identified by Harris corner detection technology, and a rectangular window with corner points as the core was constructed. Then, the rectangular window features were obtained by local binary pattern (LBP) technology, and the feature vectors were determined. Then, the feature vectors were sorted lexicographically. The image tampered area was determined by the similarity of adjacent feature vectors, and the image tampered area was determined by the similarity of adjacent feature vectors. Finally, wrong mismatching points are removed by RANSAC operator. Experimental results demonstrate that the algorithm can effectively check the degree of image tampering, and has high robustness for images tampered with by complex behaviors such as rotation tampering, brightness tampering, fuzzy tampering and noise tampering. Although the copy-paste source of multi-region tampering cannot be determined, the multi-region tampering region can be identified. The method can verify the same image copy-paste tampering image effectively. [ABSTRACT FROM AUTHOR]

铺天盖地的篡改图像已成为当今社会安全隐患之一。同图复制粘贴篡改是当前篡改图像中较为频繁使 用的篡改手法，本文尝试基于LBP 算法对同图复制粘贴篡改图像进行鉴别。先将图像转为灰度图再使用低通 滤波操作将会降低检验效果的因素减少，再运用Harris 算法提取特征点，LBP 算法提取特征向量，并将特征值 匹配，最后用RANSAC 算法来消除误匹配点。结果显示，本文算法可有效检测出经过后处理操作过的同图复 制粘贴篡改图像，该方法检验同图复制粘贴篡改图像效果良好。 [ABSTRACT FROM AUTHOR]

Copyright of Forensic Science & Technology is the property of Institute of Forensic Science, Ministry of Public Security and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)