A Comparative Study of Wavelet-Based Image Fusion with a Novel Fusion Rule II

Image fusion is the process of combining information of interest in two or more images of a scene into a single highly informative image. Generally, the multi-resolution image fusion based on the wavelet transform performs better on diverse images than traditional methods. Therefore, a comparative study of wavelet-based image fusion with different wavelet families and fusion methods are far-reaching to guide people in their applications. In our previous paper, a platform of wavelet-based image fusion has been built, which includes different wavelet families, image fusion rules, and statistical evaluation criteria. Therefore, it can be used to assess, and more importantly, select appropriate method for a specific fusion task. In this paper, the platform is further extended with more advanced algorithms and objective evaluation criteria. Discussions with these algorithms with multi-focus and multi-spectral images demonstrate that our proposed fusion approach still outperforms traditional methods in most of multi-focus images and multi-spectral images, which also complies with statistical analysis by objective evaluation criteria.