Video Copy-Move Forgery Detection and Localization Based on Structural Similarity
Copy-move forgery is one of the most common types of video forgeries. To detect such forgery, a new algorithm based on structural similarity is proposed. In this algorithm, we extend structural similarity to measure the similarity between two frames of a
- PDF / 651,321 Bytes
- 14 Pages / 439.37 x 666.142 pts Page_size
- 71 Downloads / 239 Views
Video Copy-Move Forgery Detection and Localization Based on Structural Similarity Fugui Li and Tianqiang Huang
Abstract Copy-move forgery is one of the most common types of video forgeries. To detect such forgery, a new algorithm based on structural similarity is proposed. In this algorithm, we extend structural similarity to measure the similarity between two frames of a video. Since the value of similarity between duplicated frames is higher than that between the normal inter-frames, a temporal similarity measurement strategy between short sub-sequences is put forward to detect copy-move forgery. In addition, we can obtain an accurate forgery localization. Extensive experimental results evaluated on 15 videos captured by the digital camera and mobile camera in stationary and moving mode show that the precision of this algorithm can reach 99.7 % which is higher than a previous relevant study. Keywords Video forgery Structural similarity
Copy-move detection Copy-move localization
7.1 Introduction With the wide use of a variety of digital multimedia devices as well as the development of powerful video editing tools (such as Adobe Premiere Pro and Adobe After Effects, etc.), it is becoming easy for common users to edit and process videos without leaving any visual clues. When a large number of edited and forged videos appear on the video sharing sites, the news, scientific discovery
F. Li (&) T. Huang School of Mathematics and Computer Science, Fujian Normal University, Fuzhou 350007, China e-mail: [email protected] T. Huang e-mail: [email protected]
A. A. Farag et al. (eds.), Proceedings of the 3rd International Conference on Multimedia Technology (ICMT 2013), Lecture Notes in Electrical Engineering 278, DOI: 10.1007/978-3-642-41407-7_7, Springer-Verlag Berlin Heidelberg 2014
63
64
F. Li and T. Huang
and court exhibits, there is no doubt that they will have a significant adverse effects on the stability of society and the state. Therefore, digital video forensics has become a very important research issue [1]. Video forensics can be classified into two different categories: active forensics and passive forensics. For active forensics, some pre-embedded specific information which could not be perceived in the video is needed, such as digital watermark and digital signature. In this case, one can determine whether the video is tampered or not by detecting the integrity of the information. While there is no requirement on specific information for passive forensics just by analyzing some inherent properties of videos. Recently, more attention was drawn to passive forensics. For an MPEG video, it is usually resaved in MPEG format after tampering operations. In the literature, there are already different kinds of methods for detecting video forgeries in MPEG format. In [2, 3], the authors proposed methods to detect video forgeries based on double compression and double quantization. The authors of [4] proposed a feature curve to reveal the compression history of an MPEG video file with a given GOP struct
Data Loading...
