Video forgery detection and localization using normalized cross-correlation of moment features
Manish Mathai, Deepu Rajan, S. Emmanuel · 2016
Digital technology enabled tampering of digital videos much easier using sophisticated image/video editing software. As a result, the integrity of image/video content can no longer be taken for granted and a number of forensic related issues arise paving the way for many security concerns. So detection of video forgery has become a critical requirement to ensure integrity of video data. A video forgery detection and localization method based on statistical moment features and normalized cross correlation factor is proposed. The features from prediction-error array are calculated for each frame block (set of a certain number of continuous frames in the video). The normalized cross correlation of those features between duplicated frame blocks will be high as compared to other non-duplicated ones. By using calculated threshold, based on mean-squared error, the duplication is confirmed. The location of duplicated block is also found using the algorithm. Compared to existing video forgery detection results, better true positive rates are attained.