Multiple Fiducial Identification Using the Hidden Markov Model in Image Guided Radiosurgery

• Published in: 2006 Conference on Computer Vision and Pattern Recognition Workshop (CVPRW'06) p. 92
• Main Authors: Zhiping Mu, Dongshan Fu, Kuduvalli, G.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 2006
• Subjects: Computed tomography; Feature extraction; Hidden Markov models; Lung neoplasms; Magnetic resonance imaging; Medical treatment; Robust control; Tracking; Viterbi algorithm; X-ray imaging
• ISBN: 9780769526461; 0769526462
• ISSN: 2160-7508; 2160-7516
• DOI: 10.1109/CVPRW.2006.138

A multi-fiducial identification method for image guided radiotherapy and radiosurgery is presented. A modified hidden Markov model is adopted to incorporate context information. A novel algorithm, modified from the Viterbi algorithm, is introduced to identify fiducials concurrently in two orthogonal projections. This method is robust and efficient, requires a small number of control parameters, exhibits large search range, and can accommodate deformations. A simple implementation is presented as an example to verify the efficacy of the framework. Experiments were carried out using clinical images acquired during patient treatments, and very promising results were achieved and reported in the paper. The algorithm successfully identified fiducials even in very difficult cases, demonstrating the effectiveness and robustness of the proposed probabilistic framework and the concurrent search algorithm.