VERY LARGE-SCALE INTEGRATION IMPLEMENTATION OF 3D CELLULAR AUTOMATON

KENSAL CONSULTING IS A SMALL BUSINESS IN ARIZONA WHICH HAS JUST BEEN AWARDED A UNITED STATES PATENT (NO: 4,641,351) DATED FEBRUARY 3, 1987 ENTITLED LOGICAL TRANSFORM IMAGE PROCESSOR. THE ARCHITECTURE OF THIS PATENTED PROCESSOR IS BASED ON A NOVEL CONCEPT IN THREE-DIMENSIONAL CELLULAR LOGIC. THE PURPOSE OF THIS PARTICULAR ARCHITECTURE IS TO PROVIDE AN EXTREMELY HIGH-SPEED, LOW-COST PARALLEL-PROCESSING SYSTEM FOR BOTH THREE-DIMENSIONAL SURFACE RENDERING FOR DISPLAY AND FOR THREE-DIMENSIONAL MORPHOLOGY TRANSFORMS FOR PATTERN RECOGNITION. SINCE MOST REAL-TIME, SURFACE-RENDERING DISPLAYS COST UPWARDS OF $100,000 AND EXECUTE SURFACE RENDERING BY GEOMETRIC TRIANGULATION OF EVERY SURFACE IN THE THREE-DIMENSIONAL STRUCTURE BEING DISPLAYED, A LOW-COST EQUIVALENT WOULD BE OFGREAT VALUE IN MAINTAINING THE COMPETITIVE POSTURE OF THE UNITED STATES GRAPHICS INDUSTRY. THE PURPOSE OF THE PROPOSED RESEARCH IS TO DETERMINE AN OPTIMUM VLSI IMPLEMENTATION OF THIS NOVEL ARCHITECTURE. DESIGN TRADE-OFFS OF THE LOOKUP TABLES USED FOR FLASH-COMPUTATION OF THE GRAYLEVEL OF EACH SURFACE VOXEL, THE THREE-DIMENSIONAL BIT-SLICE IMAGE MEMORY UNITS, AND ALL ASPECTS OF VOXEL ROUTING AND I/O WILL BE STUDIED. ALSO TO BE CONSIDERED IS THE CHOICE OF THE HOST COMPUTER (OR CONTROLLER) AS WELL AS METHODS OF OVERLAYING THE COMPUTED GRAYLEVEL DATA INTO REFRESH MEMORY. THE POSSIBILITY OF A DUAL-MODE DESIGN FOR BOTH DISPLAY AND MORPHOLOGY COMPUTATIONS WILL ALSO BE TREATED. FINALLY, A MARKET SURVEY WILL BE CONDUCTED IN ORDER TO PROPERLY MARRY THE RESEARCH ON THE PATENTED THREE-DIMENSIONAL PROCESSOR DESIGN WITH CURRENT NATIONAL AND INTERNATIONAL NEEDS IN THREE-DIMENSIONAL GRAPHICAL AND MORPHOLOGICAL COMPUTATION.