Image fusion algorithms for color and gray level images based on LCLS method and novel artificial neural network

In this paper, two neural image fusion algorithms for color and gray level images are proposed. These algorithms are based on a linearly constrained least square (LCLS) method and a novel projection recurrent artificial neural network. The theoretical aspects of the model are based on KKT conditions and projection theorem. Compared with the existing fusion methods, the proposed algorithms do not require any analogs multiplier and their structures are simple for implementation. Existence of the unique solution, stability and global convergence of the related projection recurrent artificial neural network model are proved. Seven steps algorithms are described in detail, for implementation. Corresponding block diagram of the entire process verifies the simplicity of these algorithms. The proposed neural network is stable in the sense of Lyapunov and converges to the optimal vector solution in a few iterations. The implementation of these algorithms for both color and gray level images shows that the quality of noisy images can be enhanced efficiently.