Table 1 Selected algorithm and used models, features and extraction scheme
From: Automatic pulmonary fissure detection and lobe segmentation in CT chest images
Authors | Models | Features | Extraction scheme |
|---|---|---|---|
Algorithms using fissure appearances | |||
â–ª The surface shaped structure | â–ª Marching cubes algorithm, Laplacian smoothing and extended Gaussian image | â–ª Implicit surface fitting using Radial Basis Functions (RBF) | |
Rikxoort et al. [17] | â–ª Difference with the other texture | â–ª Trained features | â–ª Supervised filter and classier |
Wei et al. [18] | â–ª A curvilinear line in 2D slice | â–ª Line structure | â–ª Adaptive fissure sweeping and wavelet transform |
â–ª Ridge-like structure in 2D slice | â–ª Ridgeness | â–ª Thin plate splines and maximum a posteriori estimation | |
â–ª Smooth high-intensity curve in 2D slice | â–ª Intensity or ridgeness | â–ª A curve growing algorithm modeled by Bayesian network | |
Algorithms using lung, bronchus, and vessel information | |||
Zhang et al. [23] | â–ª Smooth surface | â–ª Ridgeness image | â–ª Fuzzy reasoning system |
â–ª Ridge-like structure in 2D slice | â–ª Anatomic pulmonary atlas | ||
Ukil et al. [24] | â–ª Sparseness of the vessel | â–ª Ridgeness | â–ª 3D watershed transform |
â–ª Match with bronchus tree structure | Â | â–ª Optimal surface | |
â–ª Ridge-like structure in 2D slice | Â | Â | |
â–ª The lung borders | â–ª Trained features for fissure | â–ª Supervised filter | |
â–ª Airways and fissures | Â | â–ª Registration | |
Wei et al. [28] | â–ª Different texture for fissure | â–ª Texture analysis | â–ª Dynamic programming |
â–ª Large continuous fissure surface | â–ª Projection | Â | |
â–ª Sparseness of the vessel | â–ª The original data removed blood vessel | â–ª Cost image | |
â–ª High intensity | â–ª The vasculature | â–ª Multi-dimensional interactive watershed transform | |
â–ª Match with bronchus tree structure | â–ª The bronchial tree | Â | |
â–ª Separation by surface-shaped fissure | â–ª Pulmonary fissures | Â | |
Appia et al. [31] | â–ª High intensity | â–ª The intensity | â–ª Global minimal path |
â–ª Sparseness of the vessel | â–ª Distance of the vasculature | Â | |
â–ª Smooth in 2D | â–ª Curvature in 2D | Â | |
â–ª Continuity in 3D | â–ª Continuity in 3D | Â | |
Zhou et al. [32] | â–ª Sparseness of the vessel | â–ª Bronchus segmentation | â–ª Voronoi division algorithm |
â–ª Match with bronchus tree structure | â–ª Vessel segmentation | â–ª Laplacian filter | |
â–ª Fissure appearance of line at 2D slice | Â | Â | |