Paper Info
Title
Optimizing computed tomographic angiography image segmentation using fitness based partitioning
Abstract
Computed Tomographic Angiography (CTA) has become a popular image modality for the evaluation of arteries and the detection of narrowings. For an objective and reproducible assessment of objects in CTA images, automated segmentation is very important. However, because of the complexity of CTA images it is not possible to find a single parameter setting that results in an optimal segmentation for each possible image of each possible patient. Therefore, we want to find optimal parameter settings for different CTA images. In this paper we investigate the use of Fitness Based Partitioning to find groups of images that require a similar parameter setting for the segmentation algorithm while at the same time evolving optimal parameter settings for these groups. The results show that Fitness Based Partitioning results in better image segmentation than the original default parameter solutions or a single parameter solution evolved for all images.
Year
DOI
Venue
2008
10.1007/978-3-540-78761-7_28
EvoWorkshops
Keywords
Field
DocType
automated segmentation,cta image,original default parameter solution,single parameter,better image segmentation,different cta image,computed tomographic angiography image,optimal parameter setting,single parameter solution,similar parameter,optimal segmentation,image segmentation
Computer vision,Scale-space segmentation,Segmentation,Segmentation-based object categorization,Image segmentation,Artificial intelligence,Computed tomographic angiography,Mathematics
Conference
Volume
ISSN
ISBN
4974
0302-9743
3-540-78760-7
Citations 
PageRank 
References 
1
0.37
9
Authors
9
Name
Order
Citations
PageRank
alexander eggermont110.37
Rui Li2788.10
Ernst G. P. Bovenkamp3495.73
henk a marquering4246.74
Michael T. M. Emmerich524722.74
Aad van der Lugt624825.26
Thomas Bäck762986.94
Jouke Dijkstra812616.92
Johan H. C. Reiber91767286.53