Title :
Automatic Contour Generation of 4D CT by Deformable Registration
Author :
Duan, Zhuolei ; Zhou, Fugen ; Liu, Bo ; Deng, Xiaowu ; Chen, Ming ; Huang, Xiaoyan
Author_Institution :
Sch. of Astronaut., Beijing Univ. of Aeronaut. & Astronaut., Beijing
Abstract :
The purpose of this work is to develop a method of automatic contour generation by deformable registration on four-dimensional CT data. Firstly, a region of interest was manually delineated by a physician on a reference phase. Then deformable registration was implemented to acquire the spatial relationship between different phases. The contour on the reference phase was mapped into all other phases by the output transformation of registration. The improved demons algorithm adds symmetric gradient to the calculation of original deformable force and has smaller registration error. The accuracy of registration plays an important role to the transforming image contours. Application of this approach to 4D CT images shows that displacements of geometrical center of delineated and generated contour are within 2 mm and the maximum volume difference is just 1.5 ml. The level of accuracy is acceptable for most radiotherapy applications. The robustness and accuracy of this approach show that it is a valuable tool for 4D radiation treatment planning.
Keywords :
computerised tomography; edge detection; geometry; gradient methods; image registration; medical image processing; radiation therapy; 4D CT; 4D radiation treatment planning; automatic contour generation; deformable registration; demons algorithm; geometrical center; image contour transformation; radiotherapy applications; symmetric gradient; Aerospace industry; Biomedical applications of radiation; Cancer; Computational intelligence; Computed tomography; Computer industry; Conferences; Feature extraction; Image registration; Oncology; 4D CT; contour generation; deformable registration;
Conference_Titel :
Computational Intelligence and Industrial Application, 2008. PACIIA '08. Pacific-Asia Workshop on
Conference_Location :
Wuhan
Print_ISBN :
978-0-7695-3490-9
DOI :
10.1109/PACIIA.2008.94