Robust ellipse detection with Gaussian mixture models

The Euclidian distance between Gaussian Mixtures has been shown to be robust to perform point set registration (Jian and Vemuri, 2011). We propose to extend this idea for robustly matching a family of shapes (ellipses). Optimisation is performed with an annealing strategy, and the search for occurre...

Full description

Saved in:
Bibliographic Details
Published in:Pattern recognition 2016-10, Vol.58, p.12-26
Main Authors: Arellano, Claudia, Dahyot, Rozenn
Format: Article
Language:English
Subjects:
Benchmarking
Ellipse detection
Ellipses
Gaussian
GMM
L2 distance
Optimization
Parameter estimation
Pattern recognition
State of the art
Strategy
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The Euclidian distance between Gaussian Mixtures has been shown to be robust to perform point set registration (Jian and Vemuri, 2011). We propose to extend this idea for robustly matching a family of shapes (ellipses). Optimisation is performed with an annealing strategy, and the search for occurrences is repeated several times to detect multiple instances of the shape of interest. We compare experimentally our approach to other state-of-the-art techniques on a benchmark database for ellipses, and demonstrate the good performance of our approach. •We extend the framework based on L2 to estimate a parametric family of curves.•We propose a non-isotropic and multidimensional modeling for the density functions.•We propose a method for detecting multiple instances of an ellipse.
ISSN:0031-3203
1873-5142
DOI:10.1016/j.patcog.2016.01.017