DocumentCode
1877771
Title
FSM: A new finite sphere method for modeling 3D geometry of the prostate
Author
Zaim, Amjad
Author_Institution
Intell. & Bioinf. Comput. Sci. Dept., Univ. of Texas at Brownsville, Brownsville, TX
fYear
2008
fDate
12-15 Oct. 2008
Firstpage
2956
Lastpage
2959
Abstract
Modeling the irregular three dimensional geometry of the prostate in 3D TRUS is an important task useful in a variety of clinical procedures of the prostate such as visualization, prostate volume estimation, treatment planning. Traditionally, the prostate volumetric shape has been often modeled with primitive shapes that oversimplify its complex and deformable structure. In this paper, we present a new model, we call it Finite Sphere Model or FSM, which is composed of variable-size spheres packed inside the prostate volume. This robust method accommodates different prostate shapes and does not need prior knowledge or shape approximation. First, the prostate region is extracted and isolated from the TRUS image. The algorithm applies morphological processing in order to obtain the medial axis of the prostate. Along the medial axis, the prostate space is searched for the optimal number of spheres that can be fitted inside its volume. Finally, small patches left by the algorithm are detected by masking the image with smaller discs until the filling criteria is satisfied. A number of experiments are used to verify the efficiency of the developed model.
Keywords
biomedical imaging; cancer; DCTmod2 features; PCA-based classification; discrete pseudo 2-dimensional hidden Markov model; face recognition; facial feature hypotheses extraction; invariant curvature features; iterative closest point alignment; multistep alignment scheme; range images; vector quantization; Cancer; Computational geometry; Filling; Image segmentation; Robustness; Shape; Skeleton; Solid modeling; Ultrasonic imaging; Visualization; Finite Sphere Modeling; Mathematical Morphology; Transrectal Ultrasound; Visualization;
fLanguage
English
Publisher
ieee
Conference_Titel
Image Processing, 2008. ICIP 2008. 15th IEEE International Conference on
Conference_Location
San Diego, CA
ISSN
1522-4880
Print_ISBN
978-1-4244-1765-0
Electronic_ISBN
1522-4880
Type
conf
DOI
10.1109/ICIP.2008.4712415
Filename
4712415
Link To Document