Registration and Fusion with Mutual Information for
Information Preserved Multimodal Visualization

Augusto Valente
Profa. Wu, Shin-Ting
DCA - FEEC - UNICAMP
 

This work concerns Valente's Master's thesis in Electrical Engineering.
It is part of the The Manipulation Toolkit (MTK) project.

(a)(b)(c)(d)
(click to expand)
a) CT and b) MR images of a patient and their automatic c) registration and d) fusion (CT and MR voxels in on red and green channels respectively).

During diagnosis or surgical planning, visualizing different images in a combined form can be helpful in assessing multiple modalities. To accomplish that, the images must be spatially aligned (registered) and their intensities must be combined (fused), preferably in a way that each modality is still identifiable.

We propose the use of techniques based on Information Theory for each step available in literature to better integrate them in a common framework.

Contents

    Comming soong.

References

[1] A. Collignon et al., Automated multi-modality image registration based on information theory. Kluwer, 1995.
[2] R. Bramon et al., Multimodal data fusion based on mutual information. IEEE TVCG, vol. 99, 2011. doi:10.1109/TVCG.2011.280