Spine Image Segmentation Method Based On WFCM And RSF Active Coutour Model

Medical image processing plays an important role in clinical diagnosis and is a hot issue in the field of image processing.Medical imaging technology offers a lot of help to the doctor's treatment and the patient's rehabilitation.Image segmentation occupies very important position in medical image analysis,which substitutes manual segmentation and provides basis for 3D reconstruction.CT equipment is generally used to obtain medical images of the spine due to the higher density of the spine than the other human tissues,but the existing fuzzy effect of CT image brings huge challenge to computerized automatic segmentation.The study of spinal CT image segmentation algorithm is imminent since spinal CT image is the medical image with uneven gray level.In this paper,segmentation method based on WFCM and RSF model was adopted.At first,the original images were preprocessed effectively to make the images clearer,contrast more intense and outline more obvious in the final segmentation algorithm.After preprocessing,WFCM algorithm was applied to cluster the grayscale of spinal CT image,in the process of minimizing the objective function J,the clustering center,feature weights and dividing matrix were iterative updated until meeting the stop condition,thus far,the target and background are regarded to be ready classified and the final clustering results are obtained.WFCM coarse segmentation result was chosen as the initialization zero level set curve of RSF model,because RSF model is very sensitive to the position of initialization zero level set curve.The initialization zero level set curve evolved along with the evolution of the energy function,when the energy function reaches minimum and evolution curve reaches target curve,the final segmentation result is obtained.Finally,algorithm was verified effectively with respect of the segmentation effect,segmentation accuracy and split time in this paper.Segmentation of three different patient's 100 spinal CT images was conducted through the MATLAB platform,the text algorithm,C-V algorithm,results of doctor's manual segmentation,segmentation precision and speed of segmentation was compared,and the results proved that WFCM and RSF model algorithm proposed by text algorithm were superior to the C-V algorithm either on segmentation accuracy or on the segmentation time.Algorithm has good performance and provides theoretical basis for the computerized automatic medical image segmentation and 3D reconstruction of the spine model.