Fast image fusion based on alternating direction algorithms

In this paper, a fast image fusion approach is explored. In particular, we focus on improve the computational speed of a sparsity-based fusion framework in image fusion, where alternating direction algorithms is sought to recover sparse coefficient from high-dimensional original images. The proposed approach first compresses the sensing data by random projection and then obtains sparse coefficients on compressed samples by a fast sparse representation optimization problem based on alternating direction algorithms. Secondly, the fusion coefficients are combined with the fusion impact factor. Finally, the fused image is reconstructed from the combined sparse coefficients. We conduct extensive experiments to validate that the fusion quality of SR-RP-ADM performs better; SR-RP-ADM can effectively improve the computational speed and also can make dynamic adjustment of the compression ratio to achieve the balance between fusion quality, cost, and computing time.