Performance Analysis Image Watermarking using Discrete Cosine Transforms

Autor: Arsyad Ramadhan Darlis, Lita Lidyawati, Tio Hanif Yanuary, Lucia Jambola
Rok vydání: 2019
Předmět:
Zdroj: 2019 IEEE International Conference on Signals and Systems (ICSigSys).
DOI: 10.1109/icsigsys.2019.8811092
Popis: Digital images, video, and audio have been revolutionized in the way they can be captured, stored, transmitted, and manipulated. Therefore, the idea of using a digital watermark to detect and trace copyright violations has stimulated signifacantly. In this research will evaluate the robustness of watermark embedding with some attacks and extracting. Several functions were used to qualify the robustness of watermark algorithm, such as Peak Signal to Noise Ratio (PSNR) and Structural Similarity (SSIM). The measurements involve 3 types of watermark image, as follows $\mathbf{16\ \mathrm{x}\ 16}$ pixels, $\mathbf{32\ \mathrm{x}\ 32}$ pixels, and $\mathbf{64\ \mathrm{x}\ 64}$ pixels with the value of scaling factor (SF) and quality factor (K) of 10, 30, and 50. The simulation shows that in the condition of SF and K for 50, the watermark image of $\mathbf{64\ \mathrm{x}\ 64}$ have the highest PSNR 51.868 dB and the highest SSIM 0.983. From these results prove that the size of the watermark image could provide the better resistance to attack.
Databáze: OpenAIRE