Medical CT image preprocessing based on edge detection and spline fitting

According to the characteristic that the medical CT image data are area samplings, a new preprocessing approach for CT image is presented in this paper. It firstly constructs mapping relationship between area sampling and point sampling, then detect the image edge with Sobel operator, and divide the CT image into different regions based on human tissue feature, and finally, in each divided region, establishes fitting equations with cubic spline, so as to transform area sampling values to more accurate approximations of point sampling values. Using the new transformed values to reconstruct 3-D medical image can improve reconstruction quality. Experiments show that the 3-D reconstructed image based on CT data preprocessed by the new approach can have more detail information.