Coherence restricted StOMP and its application in image fusion

This paper proposes an image fusion method based on an improved Stagewise OMP algorithm (ApStOMP), which is more efficient than StOMP in finding the sparest solution of large-scale underdetermined problem. Multiple atoms can be selected at each stage, and only a fixed number of stages are required. Unlike StOMP, we restrict the mutual coherence of selected atoms to be as low as possible at each stage, and atoms with high coherence are excluded. In this way, we can obtain a more accurate estimated support set than with StOMP. The advantages of the proposed fusion method are demonstrated experimentally with different groups of pre-registered source images, for which we can carry out image denoising and fusion simultaneously. The experimental results show that the performance of the proposed method is competitive with other methods in terms of several objective fusion metrics, as well as in visual quality.