Lightweight Digital Watermarking Method for Videos Forgeries Detection

Abstract

With the advancement of the digital video and digital image editing tools, this increases the difficulty for humans to identify visually the authentic video from the forged copy. This needs a powerful method that investigates various video from tempering attempts. There are several previous types of research, which propose methods for video forgery detection. These works suffer from consuming the memory and much-comparing power. The consequence is increasing the execution time. This research proposes a lightweight method; this means reduce memory usage and execution time for detecting forgery in video streams. The method based on using mathematical schema structure to explore various forgeries type on streaming videos. Experiments have been conducted on a dataset of 27 videos covering different cases using local IP camera, cases like using compression algorithms and some tempering effects. The result indicated that the system achieved high relevant measures with 97% accuracy on detecting duplication on a case of lossless compression and 95% on the other compression algorithms; on the other hand, on detecting cloning forgeries our system achieved a highly relevant measure with 86% on a case of lossless compression.