An Improved Video Keyframe Detection Technique Leads to Video Authentication
Bhagyashri S. Kapre, A M Rajurkar, Guru D. S. · International Journal of Engineering Trends and Technology · 2023
Codec standard H.264/AVC compression has recently become more popular. Secure sharing of these videos is needed in applications like surveillance, remote sensing, telemedicine, and medical videos. Many times videos get edited, copied, and illegally distributed during transmission over public networks. Therefore, research on video watermarking based on this video compression standard has become very popular. In this research, a blind keyframe-based H.264/AVC video authentication technique is proposed. Embedding a watermark in each video frame increases the size of the video, and as well as it can be easily extracted by the attacker; adding a watermark in all frames is a time-consuming process. To avoid these problems, we propose a novel shot boundary detection algorithm by employing Karl-Pearson Correlation Coefficient (K-PCC) and Absolute Mean Difference (AMD), and SVD is used for the selection of keyframes. The proposed technique uses DWT, DCT, and Arnold Transform to generate a Content-Based Watermark (CBW) for each keyframe, enhancing security. These extracted keyframes are used for embedding and extraction of the watermarks. Further, DCT based embedding technique is used for embedding the watermark. The experimental findings demonstrate that the proposed method outperforms in terms of imperceptibility and robustness and can successfully detect shot boundaries under various camera actions. Also, the summarized video is effectively generated using the presented SVD-based key frame selection technique.