Bone Surface Model Development Based on Point Clouds Extracted From CT Scan Images

In the present study a procedure is proposed for the development of bone surface models by using point clouds that can be extracted from CT scan images. Since the images are as multiple two dimensional sections, three methods of surface fitting are considered: ruled, skinning and global approximation methods. The required algorithms were discussed in fields of image processing and curve and surface fitting. For the purpose of further exploring the modelling requirements and results, and gaining further insights into the impacts of effective parameters, a computer program was developed. By adopting a detailed case study and analysis approach, three samples of the cattle’s bones were selected and scanned with CT scan. Similar protocols corresponding to the human body bones were used during the scanning process. Subsequently, the surface models of the sample output from the program were transferred to CAD software. Moreover, the samples were scanned with COMET5™scanner after removing the flashes surrounding the bones. It was observed that although the bone surface modelling is feasible within 0.25 and 0.75 mm accuracy range with these algorithms, skinning method works better compared to other two algorithms in terms of processing speed and increasing the ratio of data compaction. The use of control points balance algorithm and smoothing the contours, used in this paper, will greatly improve the performance of the program as well.