Active Tampering Detection for Image Authentication using Hadamard Tracing

Image processing software that is growing rapidly makes it easier for someone to process an image for the better. However, this convenience also has a bad impact, especially if it is used in plagiarism or manipulation activities. Image authentication methods, especially those with active detection are widely developed to overcome these problems. One of the most popular methods is the SVD-based because it can generate important properties from an image. Nonetheless, most SVD-based methods only make use of the singular matrix S and ignore the other two orthogonal matrices. Therefore, this study proposes the use of Hadamard blocks to generate a new S matrix through a simpler operation. The experimental results show that the use of Hadamard blocks has the same detection capabilities and image quality as the previous method. In terms of computational time, the proposed method has a faster computational time with the highest speed improvement of 21% and 15% in the embedding and authentication process respectively. The proposed method can work faster while maintaining the level of imperceptibility and detection.