Automatic Detection of Object-Based Video Forgery Using Various Groups of Pictures (GOP)

Abstract

In recent years, there has been a lot of interest in detecting object-based video forgeries. There has been a lack of satisfactory performance with object-based forgery detectors until recently since a majority of them are still based on handcrafted features. There has been a great deal of interest in passive video forensics in recent years. Forgery of video encoded with advanced codec frameworks remains one of the biggest challenges in object-based forgery research. An object-based forgery detection approach is presented in this paper. To evaluate the proposed method, a derived test dataset of variable video lengths and frame sizes is also used in addition to the SYSU-OBJFORG dataset. This process's efficacy is verified by comparing its results with other methods. When tested on datasets with degraded-quality videos, the proposed framework performed better in real-life scenarios.