Abstract:
This work focuses on the challenges of 3D surface registration in MIS applications. A novel dataset has been created based on CT scans of a knee phantom and pre-operative...Show MoreMetadata
Abstract:
This work focuses on the challenges of 3D surface registration in MIS applications. A novel dataset has been created based on CT scans of a knee phantom and pre-operative 3D models of the meniscus are extracted. Four state-of-the-art 3D registration methods are evaluated on the registration of noisy patches segmented from the models' surface to the model itself in an attempt to objectively compare their performance using ground-truth correspondences. The results of this analysis provide useful insight on the problem at hand and indicate where our focus should be steered in order to successfully tackle it.
Date of Conference: 16-18 October 2018
Date Added to IEEE Xplore: 16 December 2018
ISBN Information:
Print on Demand(PoD) ISSN: 1558-2809