Multimodal Image Alignment via Linear Mapping between Feature Modalities

Figure 1

From left to right: (a) a pair of MSI images captured from the same retina but at different spectra; (b) overlaid results generated after alignments by the algorithms based on mutual information [19], robust measurement [4], and our linear mapping, respectively; (c) a small rectangular patch chosen at a similar position from the images from left to right at (a) and (b), respectively. The white arrow in the rectangular patches points to an area bearing obvious differences in vessel alignment between different algorithms.

(a)

(b)

(c)