Bounded Rayleigh Mixture Model for Ultrasound Image Segmentation

Abstract : The finite mixture model based on the Gaussian distribution is a flexible and powerful tool to address image segmentation. However, in the case of ultrasound images, the intensity distributions are non-symmetric whereas the Gaussian distribution is symmetric. In this study, a new finite bounded Rayleigh distribution is proposed. One advantage of the proposed model is that Rayleigh distribution is non-symmetric which has ability to fit the shape of medical ultrasound data. Another advantage is that each component of the proposed model is suitable for the ultrasound image segmentation. We also apply the bounded Rayleigh mixture model in order to improve the accuracy and to reduce the computational time. Experiments show that the proposed model outperforms the state-of-art methods on time consumption and accuracy.
Type de document :
Communication dans un congrès
8th International Conference on Graphic and Image Processing, Oct 2016, Tokyo, Japan. 2017, 〈10.1117/12.2266963〉
Liste complète des métadonnées

Littérature citée [10 références]  Voir  Masquer  Télécharger

http://www.hal.inserm.fr/inserm-01426910
Contributeur : Jean-Louis Dillenseger <>
Soumis le : jeudi 5 janvier 2017 - 10:07:14
Dernière modification le : samedi 24 mars 2018 - 01:48:34
Document(s) archivé(s) le : jeudi 6 avril 2017 - 12:20:16

Fichier

2.ICGIP_2016_paper_47-1.pdf
Fichiers produits par l'(les) auteur(s)

Identifiants

Collections

Citation

Hui Bi, Hui Tang, Hua Zhong Shu, Jean-Louis Dillenseger. Bounded Rayleigh Mixture Model for Ultrasound Image Segmentation. 8th International Conference on Graphic and Image Processing, Oct 2016, Tokyo, Japan. 2017, 〈10.1117/12.2266963〉. 〈inserm-01426910〉

Partager

Métriques

Consultations de la notice

110

Téléchargements de fichiers

132