A novel video inter-frame forgery model detection scheme based on optical flow consistency

作者：Chao, Juan; Jiang, Xinghao; Sun, Tanfeng^*

来源：11th International Workshop on Digital Forensics and Watermaking, IWDW 2012, China,Shanghai, 2012-10-31 to 2012-11-03.

DOI：10.1007/978-3-642-40099-5_22

• Adaptive thresholds
• Detection methods
• Detection scheme
• Frame deletions
• frame insertion
• Inter-frame forgeries
• Model identification
• Video forgeries

In this paper, a novel video inter-frame forgery detection scheme based on optical flow consistency is proposed. It is based on the finding that inter-frame forgery will disturb the optical flow consistency. This paper noticed the subtle difference between frame insertion and deletion, and proposed different detection schemes for them. A window based rough detection method and binary searching scheme are proposed to detect frame insertion forgery. Frame-to-frame optical flows and double adaptive thresholds are applied to detect frame deletion forgery. This paper not only detects video forgery, but also identifies the forgery model. Experiments show that our scheme achieves a good performance in identifying frame insertion and deletion model.

• 出版日期2013
• 单位上海交通大学