An international conference connecting people
in CAD research, education and business
Copyright (C) CAD Solutions, LLC. All rights reserved.
Proceedings of CAD'15, 2015, 64-68
Primitive Detection from a Single View Image of an Engineering Model
Arjun Raj, Indian Institute of Technology Madras
Abstract. 3D content-based image retrieval (CBIR) has gained paramount importance in the recent years widely due to its applicability in diverse fields from engineering to medicine. A key step in a CBIR system is to determine the features for all the objects in the database and for the query object. Typical features include color, texture, geometric primitives etc. A geometric primitive is defined as a curve or surface that can be characterized by an algebraic equation with a number of parameters. Examples of primitives typically used in creating a 3D model are spheres, prismatic boxes, cylinders, cones etc (examples for 2D primitives include lines, circles, ellipses etc.). By performing basic boolean operations on these primitives, complex engineering objects can be constructed. In the context of engineering objects, it is then apt to consider the primitives and their geometrical and topological characteristics as the high level content. It is to be noted that, with the advent of NURBS, solid models can be formed out of such representation as well, which are not considered in this paper.
Keywords. CAD/FEA transfer, CAD model simplification, Machine learning