VIDEO DUPLICATE AND FORGED FRAME DETECTION USING TWO-FOLD OPTICAL FLOW AND CORRELATION

Electronics and Communication Engineering

Video compression technologies were frequently employed to minimize the enormous size of video data. However, because of lossy reduction, it also generates unpleasant visual artifacts. Compression methods like H.264 and HEVC were used to minimize the memory space and bandwidth due to the growing volume of video data available through the Internet. Many compression artifact elimination methods have been presented to produce artifact-free pictures in the previous decades to acquire good quality images/videos on the decoding side. The research uses a two-fold video authenticating system that detects and locates both frame replication and frame forging in videos. A Lucas-Kanade-based Optical flow detection technique is designed for compression as well as artifact minimization to enhance the performance of compressed videos. The technique divides each video outline into suspicious and accurate parts initially. As a result, every part records an optical stream coefficient. During video forgery identification, feature extraction, matching, and sorting are also employed. The efficiency of the proposed technique is demonstrated by significant test findings on the Mpeg4 standard database. Manually constructed filtering and sparse coding-based approaches have already been suggested as solutions to resolve such problems.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.