Comparison of robust watermarking techniques using complex wavelet tranform.

Faryal Khanum Dotani; Muhammad Naeem Arbab; Syed Waqar Shah; Mohammad Inayatullah Babar

This paper compares the different techniques used for watermarking to obtain the most robust watermark offering various options to resist signal processing attacks. A watermark has to be strong enough to withstand the attacks. Wavelets have been used for this reason to obtain the best results. Low frequency watermarks can be used in some schemes and such schemes can be directly affected by modification of histograms. Water marks with different ranges of frequencies ranging from middle to high are extremely robust to noise and deformation having non-linear nature. Discrete wavelet transform (DWT) is an effective way of robust watermarking. Dual tree complex wavelet transform is another possible solution for robust watermarking. Complex wavelet transform uses filters just like any real time filers having selectivity highly directional in nature. Hence for real time watermarking solution, DT CWT can be an ideal solution with reduced complexity. In the process of embedding watermark, a Human visual system also plays an important role so is also considered. The strength of the watermark to be imperceptible or perceptible, as well as the quality of the images is an important criterion for embedding. DT CWT yield results taking into consideration the HVS. Comparison of the watermarking process for DT CWT and DWT to obtain a robust watermark is outlined in this paper. The results of DT CWT technique are shown to be better. In general it is a better technique for all type of images but proves to be best for low frequency, high information content images.