A.L. Abbott

Computed tomograph has extended outside of the medical imaging arena where there are now many examples that can benefit for high speed image processing algorithms to detect explosives in luggage and detect cracks in structual materials to mention only a few industrial applications. With high speed neural networks we propose to design a system that automatically detects defects or structures of interests in a 3D visualizaiton environment. Particular interest will devoted to studying the use of voxel volume visualization tools that show potential for real-time industrial applications. Current reserach show that the neural-net approach can take advantage of parallel computing architectures and hence take advantage of recent parallel computing systems on campus: Intel Paragon, IBM-SP2, and a SGI Power Challenge. a SGI Power Challenge.-SP2, and