Generation of Convex Hull Using Neural Network Technique Nayyar Ashish Kumar*, Sapra Varun** *IITM, New Delhi **JIMS, New Delhi Online published on 27 July, 2015. Abstract The convex hull problem has had a long history going back to the beginning of computational geometry and has been an intensively studied subject even up to the present day. The computation of the convexhull of a finite set of points, particularly on the plane, has been studied extensively and has wide applications in pattern recognition, image processing, cluster analysis, statistics, robust estimation, operations research, computer graphics, robotics, shape analysis, and several other fields. A convexhull based shape representation is suitable for classification and recognition of irregular objects because it is invariant with respect to coordinate rotation, translation, and scaling. Since 1970s, the problem of convex-hull computation has been an interesting area of research. As a result, a wide variety of algorithms are available in the literature to solve this problem. Early papers dealt primarily with the planar case d = 2. These wide varieties of algorithms were classified as (a) computing exact convex hull and (b) computing approximate convex hull. They were also known as sequential and parallel. Sequential means using a single processor for the computation and parallel means using multiple processors for the purpose of computation. This research aims in finding the solution of the convex hull problem using a neural network technique. The convex hull problem is the problem of computing the convex hull of S and reporting the points on the convex hull in the order in which they appear on the hull where S = {S[0], …, S[n – 1]} be a set of n distinct points in the Euclidean plane. This problem has been solved efficiently using standard methods but our research is confined to find a solution of the problem using neural network. Because convex hull is basically a pattern recognition problem and neural networks are good at pattern recognition. Top Keywords Neural, Convex, Hull, Euclidean, Computation, etc. Top |