Advances in the use of neurophysiologycally-based fusion for visualization and pattern recognition of medical imagery

The ever increasing number of image modalities available to doctors for diagnosis purposes has established an important need to develop techniques that support work-load reduction and information maximization. To this end, we have improved on an image fusion architecture first developed for night vision applications. This technique, presented at Fusion 2002, utilizes 3D operators to combine volumetric image sets while maximizing information content. In our approach, we have combined the use of image fusion and user- defined pattern recognition within a 3D human-computer interface. Here, we present our latest advances towards enhancing information visualization and supporting pattern recognition. We also report on results of applying image fusion across a variety of patient cases. Finally, we have also begun the assessment of pattern recognition based on 2D vs. 3D fused image features. Initial results indicate an advantage to fusing imagery across all three dimensions so as to take advantage of the volumetric information available in medical data sets. A description of the system and a number of examples will serve to illustrate our ongoing results.