﻿Shape Detection

Sven Oesau, Yannick Verdie, Clément Jamin, Pierre Alliez, Florent Lafarge, Simon Giraudot, Thien Hoang, and Dmitry Anisimov

This CGAL package implements the Efficient RANSAC (RANdom SAmple Consensus) algorithm
for detecting shapes in a point set with unoriented normals. Five canonical shapes
can be detected: planes, spheres, cylinders, cones, and tori.
Additional shapes can be detected, given a custom shape class by the user.

In addition to the RANSAC algorithm, the generic Region Growing algorithm is implemented.
This algorithm can detect shapes in a set of arbitrary items. In particular,
CGAL provides several instances of the algorithm: detecting lines and circles in a 2D point set,
detecting planes, spheres, and cylinders in a 3D point set, detecting planes on a polygon mesh, detecting lines
in a 2D/3D segment set, and detecting lines on a 2D/3D polyline.
Other custom instances can be easily added by the user.
