An integrated approach for reconstructing a surface model of the proximal femur from sparse input data and a multi-resolution point distribution model: an in vitro study.

BACKGROUND

Accurate reconstruction of a patient-specific surface model of the proximal femur from preoperatively or intraoperatively available sparse data plays an important role in planning and supporting various computer-assisted surgical procedures.

METHODS

In this paper, we present an integrated approach using a multi-resolution point distribution model (MR-PDM) to reconstruct a patient-specific surface model of the proximal femur from sparse input data, which may consist of sparse point data or a limited number of calibrated X-ray images. Depending on the modality of the input data, our approach chooses different PDMs. When 3D sparse points are used, which may be obtained intraoperatively via a pointer-based digitization or from a calibrated ultrasound, a fine level point distribution model (FL-PDM) is used in the reconstruction process. In contrast, when calibrated X-ray images are used, which may be obtained preoperatively or intraoperatively, a coarse level point distribution model (CL-PDM) will be used.

RESULTS

The present approach was verified on 31 femurs. Three different types of input data, i.e., sparse points, calibrated fluoroscopic images, and calibrated X-ray radiographs, were used in our experiments to reconstruct a surface model of the associated bone. Our experimental results demonstrate promising accuracy of the present approach.

CONCLUSIONS

A multi-resolution point distribution model facilitate the reconstruction of a patient-specific surface model of the proximal femur from sparse input data.