Multimodal image fusion between CBCT and SPECT: application to body extremities

The multimodal image fusion of biomedical images consists in creating a single image from images obtained with different modalities on the same subject. This method allows additional analysis and diagnosis, as it extracts the information content of two different images, synthesizing it in the merged image it produces. This overview focuses on image fusion applied to upper and lower body extremities, such as foot, ankle, hand and wrist, in order to help physicians in reducing equivocal interpretations of abnormal bone conditions due to trauma, inflammation, infection, degeneration or tumor. In this work, Cone Beam Computed Tomography (CBCT) is selected as the imaging modality for investigating bone morphology, but it does not give enough information about metabolic functionality associated to those structures. For this reason, Single-Photon Emission Computed Tomography (SPECT) can be used to obtain images which have low specificity, but high sensitivity, compensating this functional information not provided by Computed Tomography. Since images are acquired in different time and conditions, a preliminary co-registration is necessary: several methods are analyzed, evaluating if a rigid transformation is sufficient, or a non-rigid transformation is required for deformable tissue alignment. The similarity method to be optimized must be chosen according to images characteristics and degree of comparison. A multi-resolution analysis is also evaluated for optimization algorithm strengthening.