ISBI'10 Proceedings of the 2010 IEEE international conference on Biomedical imaging: from nano to Macro
Pattern Recognition and Image Analysis
| Hi-index | 0.00 |
In this paper, we present a relational-tubular (ReTu )deformable model for segmenting a complex and theentire tubular network structure with branches in closeproximity of each other. Specifically, we incorporate apriori knowledge of the target anatomy structure as wellas the spatial relationship between branches to reducepossible segmentation errors due to the effects of a varietyof imaging artifacts and noise. To get more robustdescription of the data properties than a simple 3D edgemap, a new data energy functional is proposed based ontesting the volumetric density within the model cross-sections. The deformation process is formulated as a two-stage procedure:tubular medial axis deformation andtubular surface deformation. The efficiency of thisapproach is demonstrated by our experiments which showthat satisfactory quantifications of the entire zebrafishvasculature recorded from the fluorescence confocalmicroscope. The experiments also demonstrate therobustness of our deformable model in the presence ofcomplex tissue structure that adhered to the vesselbranches.