Sequential & Patch Analysis Base Video Forgery Detection System Using Deep Learning

Abstract

Visual monitoring has become a crucial asset for overseeing and guaranteeing safety .It's fascinating to see how security applications have turn becoming a crucial component of many organizations and locations. However, there's always a risk of surveillance footage getting tampered with, which can have serious consequences. The worst part is, it's not that difficult to doctor these videos by removing objects taken from the scene, leaving no trace behind. This poses a significant challenge in ensuring the reliability of video content. Investigators examining a number of approaches to tackle this  problem,  and one promising solution is is founded upon equential and patch analyses.Similarly, video sequences can  be modeled as a mixture of normal and anomalous patches to detect and localize any tampering. The approach also involves visualizing the movement of removed objects using anomalous patches, which can help in precisely identifying the forged regions in the video. The best part is that this kind of approach is efficient and .The research results have been quite promising, and this approach has shown great potential in detecting video forgery. With the growing importance of video surveillance in ensuring security, it's crucial to have reliable methods to detect tampering and ensure the authenticity of video content.