Fast Adapting Mixture Parameters Schemes forProbability Density Di erence-based DeformableModel

Auteurs :  Aicha Baya Goumeidane, Nafaa Nacereddine
Année : 2018
Domaine : Electronique
Type : Communication
Conférence: 23rd Iberoamerican Congress on Pattern Recognition (CIARP 2018)
Résumé en PDF :  (résumé en pdf)
Fulltext en PDF : 
Mots clés : Active contour, Adaptive mixture, GMM parameters up-dates.

Résumé : 

This paper presents a new region-driven active contour usingthe pdf di erence to evolve. The pdf estimation is done via a new andfast Gaussian mixture model (GMM) parameters updating scheme. Theexperiments performed on synthetic and X-ray images have shown notonly an accurate contour delineation but also outstanding performancein terms of execution speed compared to the GMM estimation based onEM algorithm and to non-parametric pdf estimations.