Medical Applications of NDT Data Fusion

The fusion of different sources of information has always been a component of medical practice. The complexity of biological phenomena is such that they cannot be explained with a single exploration. The complementary nature of the available exploration techniques (modalities) helps the physician in refining his diagnosis, preparing or performing therapeutic procedures. In this respect it is interesting to note that the development of new medical modalities has not led to the replacement of former ones, and that obviously there is no single modality providing the clinician with all possible sources of information. Before the development of computerised registration tools, data fusion involved pure mental matching of the data sets based on well-known common structures, the matching of the rest of the data being mentally interpolated from this initial step. The parallel emergence of new digital medical imaging devices, communication networks and powerful workstations has made it possible not only to display images but also to transfer and process them. Image processing methods have recently been developed to perform direct, automatic data matching. These methods define the multimodal data fusion topic. They have modified the way multimodal matching is performed as well as the way multimodal information is used; moving from a mental to a computer assisted fusion process. This evolution has led to a more accurate, more visual, more quantitative and therefore more objective fusion process.