作者: Guoyan Zheng , Weimin Yu
DOI: 10.7892/BORIS.76879
关键词: Aerospace electronics 、 Computer vision 、 Image registration 、 Artificial intelligence 、 Set (abstract data type) 、 B-spline 、 Reference space 、 Proximal femur 、 Domain (software engineering) 、 Iterative reconstruction
摘要: In this paper we present a new approach for personalized X-ray reconstruction of the proximal femur via non-rigid registration 3D volumetric template to 2D calibrated C-arm images. The 2D-3D is done with hierarchical two-stage strategy: global scaled rigid stage followed by regularized deformable b-spline stage. both stages, set control points uniform spacing are placed over domain and registrations driven computing updated positions these points, which then allows accurately register reference space Comprehensive experiments on simulated images, images cadaveric femurs clinical datasets designed conducted evaluate performance proposed approach. Quantitative qualitative evaluation results given, demonstrate efficacy