DocumentCode :
454906
Title :
Hybrid Skeleton Graph Analysis of Disordered Porous Media. Application to Trabecular Bone
Author :
Aufort, G. ; Jennane, R. ; Harba, R. ; Benhamou, C.L.
Author_Institution :
LESI, Univ. d´´Orleans
Volume :
2
fYear :
2006
fDate :
14-19 May 2006
Abstract :
In this paper, we present a new method for modeling disordered porous media. Recent work has shown that the line skeleton is a powerful tool for 3D structure characterization. However, as it generates only 1D curves, the geometry of the material is sometimes excessively approximated. Our algorithm is an improved solution as it considers the local shape of each element that composes the structure of the medium. It consists in an efficient combination of curve and surface thinning techniques. Features extracted from the new skeleton contain significant topological and morphological information. A clinical study carried out on trabecular bone samples demonstrates the ability of our method to discriminate between 2 different populations
Keywords :
bone; feature extraction; medical image processing; 1D curves; 3D structure characterization; disordered porous media modeling; feature extraction; hybrid skeleton graph analysis; line skeleton; surface thinning techniques; trabecular bone; Cancellous bone; Data mining; Feature extraction; Geometry; Shape; Skeleton; Slabs; Surface morphology; Switches; Topology;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Acoustics, Speech and Signal Processing, 2006. ICASSP 2006 Proceedings. 2006 IEEE International Conference on
Conference_Location :
Toulouse
ISSN :
1520-6149
Print_ISBN :
1-4244-0469-X
Type :
conf
DOI :
10.1109/ICASSP.2006.1660459
Filename :
1660459
Link To Document :
بازگشت