Akademik Veri Yönetim Sistemi
TURKISH JOURNAL OF ELECTRICAL ENGINEERING AND COMPUTER SCIENCES, cilt.28, sa.6, ss.3235-3252, 2020 (SCI-Expanded)
This paper evaluates and analyses the discrete wavelet transform (DWT) frequency bands for embedding and extracting of the biometric data using DWT single level and multilevel watermarking approach with and without the use of alpha blending approach. In addition, singular value decomposition (SVD) combined with DWT is used to embed and extract the watermark image. The performance of compression and decompression approaches has been analyzed to examine the robustness and to check whether the compression function does destroy the integrity of the watermarked image. We investigate the proposed approach to understand how robust the watermarked on different sub-band is against the image processing and geometric attacks. The experimental results show that the DWT multilevel and SVD watermarking approach is more robust than the other watermarking methods implemented in this paper. The DWT bands dominate the approximate coefficient band (LL) in high-intensity alpha value. When the intensity is reduced, the coefficient standout to be very robust against the remaining bands.