An application of compressive sensing for image fusion

Abstract

Compressive sensing (CS) has inspired significant interests because of its compressive capability and lack of complexity on the sensor side. This paper introduces a novel framework of image fusion based on the CS principle. First, we present a study of three sampling patterns and investigate their performance on CS reconstruction. We then propose a novel image fusion algorithm by using an improved sampling pattern. Finally, the CS-based image fusion approach is applied to various image modalities and evaluated both visually and in terms of fusion quality metrics. The simulations demonstrate that CS-based image fusion has a number of perceived advantages in comparison with image fusion in the multiresolution (MR) domain, providing a truly different and more advanced way for fusing multimodality images.

Keywords:

• compressive sensing
• CS-based image fusion
• multiresolution image fusion

2010 AMS Subject Classification :

• 68U10

Acknowledgements

This work is partially funded by the NCET of MOE, China Scholar Council and the SRF for ROCS, SEM.

Notes

An image can be vectorized into a long one-dimensional vector.