An Unsupervised Algorithm for Detecting Good Continuation in Dot Patterns
José Lezama, Gregory Randall, Jean-Michel Morel, Rafael Grompone von Gioi
⚠ This is a preprint. It may change before it is accepted for publication.


In this article we describe an algorithm for the automatic detection of perceptually relevant configurations of 'good continuation' of points in 2D point patterns. The algorithm is based on the 'a contrario' detection theory and on the assumption that 'good continuation' of points are locally quasi-symmetric. The algorithm has only one critical parameter, which controls the number of false detections.