Investigation of Feature-Based Nonrigid Image Registration Using Gaussian Process
- 49 Downloads
For a wide range of clinical applications, such as adaptive treatment planning or intraoperative image update, feature-based deformable registration (FDR) approaches are widely employed because of their simplicity and low computational complexity. FDR algorithms estimate a dense displacement field by interpolating a sparse field, which is given by the established correspondence between selected features. In this paper, we consider the deformation field as a Gaussian Process (GP), whereas the selected features are regarded as prior information on the valid deformations. Using GP, we are able to estimate the both dense displacement field and a corresponding uncertainty map at once. Furthermore, we evaluated the performance of different hyperparameter settings for squared exponential kernels with synthetic, phantom and clinical data respectively. The quantitative comparison shows, GP-based interpolation has performance on par with state-of-the-art B-spline interpolation. The greatest clinical benefit of GP-based interpolation is that it gives a reliable estimate of the mathematical uncertainty of the calculated dense displacement map.
Unable to display preview. Download preview PDF.
- 1.Rueckert D, Sonoda LI, Hayes C, et al. Nonrigid registration using free-form deformations: application to breast MR images. IEEE Trans Med Imaging. 1999 Aug;18(8):712-721.Google Scholar
- 2.Bayer S, Zhai Z, Strumia M, et al. Registration of vascular structures using a hybrid mixture model. Int J Comput Assist Radiol Surg. 2019 June;14.Google Scholar
- 3.Bookstein FL. Principal warps: thin-plate splines and the decomposition of deformations. IEEE Trans Pattern Anal Mach Intell. 1989 Jun;11(6):567-585.Google Scholar
- 4.Reinertsen I, Descoteaux M, Siddiqi K, et al. Validation of vessel-based registration for correction of brain shift. Med Img Anal. 2007;11(4):374–388.Google Scholar
- 5.Rasmussen CE, Williams CKI. Gaussian processes for machine learning. MIT Press; 2006.Google Scholar
- 6.Wachinger C, Golland P, Reuter M, et al. Gaussian process interpolation for uncertainty estimation in image registration. In: Proc MICCAI; 2014. p. 267–274.Google Scholar
- 7.Luo J, Toews M, Machado I, et al. A Feature-Driven active framework for Ultrasound-Based brain shift compensation. In: Proc MICCAI; 2018. p. 30–38.Google Scholar
- 8.Kallis K, Kreppner S, Lotter M, et al. Introduction of a hybrid treatment delivery system used for quality assurance in multi-catheter interstitial brachytherapy. Phys Med Biol. 2018 may;63(9).Google Scholar
- 9.Bishop CM. Pattern recognition and machine learning. Berlin: Springer; 2006.Google Scholar
- 10.Bayer S, Maier A, Ostermeier M, et al. Generation of synthetic image data for the evaluation of brain shift compensation methods. In: Proc CIGI; 2017. p. 10.Google Scholar
- 11.Bayer S, Wydra A, Zhong X, et al. An anthropomorphic deformable phantom for brain shift simulation. In: IEEE Nucl Sci Symp Conf Rec; 2018. p. 1–3.Google Scholar