A WATERMARKING RELATIONAL DATABASE BASED ON 5-LEVEL DWT USING GENETIC ALGORITHM & PSO

A watermarking scheme for digital images based on Particle Swarm Optimization (PSO) in the Discrete Wavelet Transform (DWT) is proposed. The watermark is inserted into the DWT sub bands which have the most important coefficients. The robustness of proposed scheme is empowered by applying the PSO. PSO optimizes the imperceptibleness of the watermark and the quality of the watermarked image which results in identifying the optimal / nearly optimal embedding positions. A series of experiments were carried out using different host images with different watermarks under different attacks. It has been shown that the approach is robust against several watermarking attacks that may meet the watermarked image. Identification of owner is cryptographically made secure and used as an embedded watermark. An improved hash partitioning approach is used that is independent of primary key of the database to secure ordering of the tuples. Strength of PSO is explored to make the technique robust, secure and imperceptible. In this paper, we have prearranged for a brief overview about the proposed hybrid model optimization in which the embedding and extracting algorithms of watermarking in discrete wavelet transform are combined with GA-PSO based optimization techniques for watermarking. The key parameters to be concentrated for this proposed model are orthogonality, symmetry and compact support which will enable the model to achieve a better watermarked media and robustness in watermarking. The watermarking technique proposed in this work may be very effective against different low-frequency an attack that demolishes the low frequency component of the image. For the implementation of this proposed work we use the Image Processing Toolbox under Matlab software.