Knot Segmentation in Noisy 3D Images of Wood

Adrien Krähenbühl 1, * Bertrand Kerautret 1 Isabelle Debled-Rennesson 1
* Auteur correspondant
1 ADAGIO - Applying Discrete Algorithms to Genomics and Imagery
LORIA - ALGO - Department of Algorithms, Computation, Image and Geometry
Abstract : Resolving a 3D segmentation problem is a common challenge in the domain of digital medical imaging. In this work, we focus on another original application domain: the 3D images of wood stem. At first sight, the nature of wood image looks easier to segment than classical medical image. However, the presence in the wood of a wet area called sapwood remains an actual challenge to perform an efficient segmentation. This paper introduces a first general solution to perform knot segmentation on wood with sapwood. The main idea of this work is to exploit the simple geometric properties of wood through an original combination of discrete connected component extractions, 2D contour detection and dominant point detection. The final segmentation algorithm is very fast and allows to extract several geometrical knot features.
Type de document :
Communication dans un congrès
Gonzalez-Diaz, Rocio and Jimenez, Maria-Jose and Medrano, Belen. 17th IAPR International Conference on Discrete Geometry for Computer Imagery - 2013, Mar 2013, Sevilla, Spain. Springer Berlin Heidelberg, 7749, pp.383-394, 2013, 〈10.1007/978-3-642-37067-0_33〉
Liste complète des métadonnées

https://hal.inria.fr/hal-00804070
Contributeur : Adrien Krähenbühl <>
Soumis le : dimanche 24 mars 2013 - 18:38:12
Dernière modification le : mardi 24 avril 2018 - 13:29:35

Identifiants

Citation

Adrien Krähenbühl, Bertrand Kerautret, Isabelle Debled-Rennesson. Knot Segmentation in Noisy 3D Images of Wood. Gonzalez-Diaz, Rocio and Jimenez, Maria-Jose and Medrano, Belen. 17th IAPR International Conference on Discrete Geometry for Computer Imagery - 2013, Mar 2013, Sevilla, Spain. Springer Berlin Heidelberg, 7749, pp.383-394, 2013, 〈10.1007/978-3-642-37067-0_33〉. 〈hal-00804070〉

Partager

Métriques

Consultations de la notice

202