An AI Approach to Computer Assisted Tomography

Computer assisted tomography (CAT) systems demandlarge amounts of time and space. In this paper, wedescribe an approachto solving the CAT problemusing several AItechniques including representation selection and pruning. Rather than backprojectintensity data onto a pixel grid or voxel lattice, wesolve the problem geometrically -- a backprojection region is stored as a convexpolygon in a planar decomposition. Analgorithm for intersecting two planar decompositions is described below. This algorithm allows us to merge intensity information frommultiple sources into a single coherent structure. Knowledgeof the task can providepruninginformation in the form of stabilized regions that neednot to be decomposed. Background Standardmedical and industrial computerassisted tomography (CAT)systems collect a series of 2-Dx-ray images an object and produce a 3-Dreconstruction of the interior structure of the object (Figure I ). The2-Dimagesare backprojected (Herman, 1980) from the image through the ...

• Type: Conference
• Year: 1998
• Where: FLAIRS
• Authors: John F. Kolen, David A. Shamma, Thomas Reichherzer, Timothy Flueharty
• Keywords: Artificial Intelligence, Computer Assisted Tomography, FLAIRS 1998, Planar Decompositions, Systems Demandlarge Amounts