Title :
A Novel Level Set Initialization Method
Author :
Liu Tao ; Zhou Qinwu ; Bian Zhengzhong ; Cao Tiesheng
Author_Institution :
Sch. of Life Sci. & Technol., Xi´an Jiaotong Univ., Xi´an, China
Abstract :
Level set method is an important numerical method for PDE image analysis, with the disadvantages of high computation cost, slow speed, and especially frequent reinitialization. The present study carefully investigates these disadvantages and proposes a new initialization method based on mathematical morphology. The new method includes smoothing scheme and reconstruction scheme, which perform the two operations of smoothing inaccurate distance function and rebuilding narrow band respectively. The method introduced in this study is not only characterized by high efficiency, but also can adapt to various topology changes. In conclusion, this method can replace the conventional methods successfully and have application in the real-time image processing.
Keywords :
image reconstruction; mathematical morphology; partial differential equations; contour distillation; image analysis; image processing; image reconstruction; image smoothing; level set initialization; mathematical morphology; partial differential equations; Biomedical imaging; Costs; Image processing; Image reconstruction; Level set; Morphology; Narrowband; Partial differential equations; Smoothing methods; Topology;
Conference_Titel :
Image and Signal Processing, 2009. CISP '09. 2nd International Congress on
Conference_Location :
Tianjin
Print_ISBN :
978-1-4244-4129-7
Electronic_ISBN :
978-1-4244-4131-0
DOI :
10.1109/CISP.2009.5304593