VIDEO WATERMARKING USING SINGULAR VALUE DECOMPOSITION ALGORITHM

• Main Author: Ginting, Gloryanson
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/75419
• Institution: Institut Teknologi Bandung

Social media is a communication platform that has become inseparable part of human life in the 21st century. Social media is generally used by users to share content in the form of both images and videos. However, social media users often commit copyright infringement, where users distribute images and videos without the permission of the copyright owners. One solution to this problem is the use of watermarks to facilitate the identification of copyrighted videos on social media. The use of steganography techniques to embed watermarks in images and videos is known as watermarking. Singular Value Decomposition (SVD) is one of the watermarking algorithms that has advantages such as visibility, resistance to attacks, and fast computation time. The results of experiments in this study found that the watermark extraction process using the SVD algorithm has relatively low quality with an average PSNR value of 16.5 db. This is because the quality of the watermark is correlated with the bitrate of videos used. The study also found that the SVD algorithm is resistant to noise, filter, and compression attacks. Furthermore, the experimental results showed that the SVD algorithm has relatively good performance with fast computation time, averaging at 7 minutes and 54 seconds. The performance of the watermark extraction algorithm is also relatively fast, with an average time of 7.597 seconds for extraction from videos without any attacks, 7.387 seconds for extraction from videos that have undergone filter attacks, 7.485 seconds for extraction from videos that have undergone noise attacks, and 7.532 seconds for extraction from videos that have undergone compression attacks.