Tampering Detection In Oral History Video Using Watermarking

The oral history videos present highly cultural and historical values. The protection of its authenticity and integrity are important. A lot of video watermarking schemes are proposed to prevent malicious attacks in recent years. An innovative algorithm is proposed in this work to resist key frames dropping and detect the subtitle tampered. Firstly, Simple Linear Iterative Clustering (SLIC) Superpixels are used to extract the natural codes of the video. Meanwhile, the targeting codes are obtained from the natural codes which are modified by the (7, 4) Hamming code error correction for identifying the watermarked frame. Secondly, the text region is located accurately by structure tensor and subtitle segmentation. Furthermore, the watermark information is composed of stroke and structural characteristic of Chinese characters. Finally, the watermark is embedded into the oral history videos based on Discrete Cosine Transform (DCT) and Singular Value Decomposition (SVD) robustly. The experimental results show the algorithm is against to several attacks including tampering texts, frames dropping and compression.