Abstract # 2996 A New Hierarchical Symmetric Optical Flow-based Image Registration Method and its Clinical Evaluation

Presenter: Michalski, Darek

< Back to topics

> Cover Page, Authors, Institutions

> Purpose/Objective

> Materials/Methods

> Results/Discussion

> Conclusions

> View tables/images/videos

The classical variational optical flow method and the resultant field of the velocities is based on a gradient fields extracted from a template image. We have extended this model through an addition of an extra virtual image incorporated into the model as a computational construct half‐way along displacements. Thus a gradient field of a scene is incorporated into the algorithm. The computation is built upon the multi‐resolution approach based on coarse‐to‐fine computation with spline pyramids. We explored the sensitivity of the calculation to gradient evaluation which is critical for this approach. Three methods were used: (i) finite difference method, (ii) B‐spline differentiation and (iii) paired/matched multi‐point filters. Registered images are obtained with the backward reconstruction. The software program was implemented with C‐programming language. CT scans of thorax used for treatment planning of lung cancer patients were employed for this study.