A new fusion image quality assessment based on edge and structure similarity

摘要

The quality assessment of fusion images is of fundamental importance for most image fusion applications. Piella proposes a quality metric for image fusion based on structural similarity but it has been proved not so good in some cases. In this paper we present a novel objective fusion image quality index which is also very easy to calculate and applicable to various image fusion processing applications. Statistics by many experiments, we find that in most cases when the edge image of fused results are closer to the addition result of the two original edge images, it can express better visual effects. So, we introduce the edge information and the corresponding addition result into the assessment and calculate the whole index based on the ideal edge structure both in part and unity. Experiments demonstrate that the new assessment maintains better consistency with human subjective perception.