Application of CAN2 to plane extraction from 3D range images

An application of CAN2 (competitive associative net 2) to plane extraction from 3D range images obtained by a LRF (laser range finder) is presented. The CAN2 basically is a neural net which learns efficient piecewise linear approximation of nonlinear functions, and in this application it is utilized for learning piecewise planner surfaces from the range image. As a result of the learning, the obtained piecewise planner surfaces are much smaller and much more than the actual planner surfaces, so that we introduce a method to gather piecewise planner surfaces for reconstructing the actual planner surfaces. We apply this method to real range images, and examine the performance and the comparative advantage to other methods.