A Ga-based Nearly Optimal Image Authentication Approach

In this paper, we present how to find nearly optimal positions for embedding authentication message by Genetic Algorithm (GA), so as to achieve high quality protected image in image authentication problem. Correlations between important DCT coefficients and user defined thresholds constitute the image authentication message. The embedding positions are simulated as chromosomes and we use GA operators, such as reproduction, crossover, and mutation to find nearly optimal embedding positions. Experimental results demonstrate that GA can improve the image quality of protected image effectively.