Shape-Preserving Fitting Used for Medical Image Pre-processing

We present a novel technique for pre-processing medical images based on the feature that medical image data obtained from CT scanner are regional sampled values. The new technique first establishes mapping relation between regional sampled values and point sampled values, then adopts cubic spline curve fitting to convert the regional sampled values to more precise approximations of point sampled values. Three-dimensional reconstruction with the converted new values can reduce reconstructed model errors, and improve reconstructed image quality. Comparison experiments are included in this paper.