Analysis of watermarking techniques with wavelet and cosine transform for imperceptibility on image

• 主要な著者: Abdullad, Maryam Jasim, Ismail, Amelia Ritahani, Abubakar, Adamu
• フォーマット: Conference or Workshop Item
• 言語: English; English; English
• 出版事項: IEEE Explore 2021
• 主題: QA75 Electronic computers. Computer science
• オンライン･アクセス: http://irep.iium.edu.my/93866/1/Analysis%20of%20Watermarking%20Techniques%20with%20Wavelet%20and%20Cosine%20Transform%20for%20Imperceptibility%20on%20Image.pdf; http://irep.iium.edu.my/93866/7/93866_Analysis_of_%20Watermarking_%20Techniques.pdf; http://irep.iium.edu.my/93866/12/93866_SCOPUS_Results_Final.pdf; http://irep.iium.edu.my/93866/; https://ieeexplore.ieee.org/abstract/document/9429305

Some medical image watermarking techniques utilize frequency transform domains to improve the performance of the watermarking technique, where the main requirement of the medical image is to maintain the visual quality. This paper analyzes the imperceptibility performance of two frequency transform techniques Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT). The analysis methodology is an experiment the embedding watermark data in the transferred domain of the medical image by image by utilizing two techniques DWT and DCT. The performance of both techniques was measured in terms of image imperceptibility and embedding capacity. The imperceptibility performance based on Peak Signal-toNoise Ratio (PSNR) and Bit Error Rate (BER) values of the watermarked medical image is 97.4129 to 104.0503 on the PSNR scale for three different medical images by DWT, while the BER range from 0.0038 to 0.0146 respectively. While under DCT, the performance under PSNR for three different medical images are 40.97, 43.15, 11.89 respectively, and BER of 0.39, 0.2995, 0.4259 respectively. This finding indicates that DWT outperformed DCT on PSNR.