Automatic Segmentation of the Body and the Spinal Canal in CT Images Based on A Priori Information

A new method was presented by using a priori information to perform automatic segmentation of the body and the spinal canal in Computed Tomography (CT) images. In this approach, the extraction of the body contour was considered as a part of the procedure in the spinal canal segmentation. The approach consists of four stages: firstly, the body contour was obtained based on the methods of thresholding and marching squares; secondly, with the usage of the body contour, a region of interest (ROI) in CT image was extracted to eliminate the affection from other tissues; thirdly, the seed pixel for fuzzy segmentation of the spinal canal was detected in the ROI image; and finally the spinal canal was segmented by the fuzzy region growing method. Normal CT studies in thorax and abdomen of real patients were used for the test. The results demonstrated the efficiency of the proposed approach in contouring the body and the spinal canal. And it also indicated that a priori information could be useful for the segmentation of organs.