tarquin@athena.mit.edu (Robert P Poole) (09/23/90)
I am writing some CAD software for the PC, and I would like to implement hidden edge removal using neural net models. Specifically, I want to implement a back face cull (for starters), which is an algorithm that computes which direction a polygon is facing and then, depending whether that polygon is facing the observer or not, draws it or not. I thought that perhaps I could simulate a neural net and, feeding it coordinate data on the three-dimensional endpoints of the line segments, train it to remove certain lines based on their placement. I would sit through the training and every time it drew a line or shouldn't (or didn't draw a line it should have), I would tell it which line segment it got wrong, etc. Anyone know which neural net model would be best, and have any ideas how it would be implemented? Thanks in advance! -- Robert P. Poole tarquin@athena.mit.edu 46 Massachusetts Avenue MIT Course VIII 311B Bexley Hall "We make Idols of our concepts, but Cambridge, MA 02139 wisdom is born of wonder."