CT Image Kidney Segmentation And Three-dimensional Reconstruction And Its Application In Percutaneous Nephrolithotomy

Urinary calculi is one of the common diseases in urology,with the proportion of kidney stones can reach more than 70%.Data statistic shows that the incidence of urinary calculi in the country is 1%~ 5%,and the southern region is up to 5%~10%.In clinic,percutaneous nephrolithotomy is mainly used to treat calculi which is above 2 cm in diameter and at renal pelvis and renal calyx.The difficulty of operation lies in the establishment of percutaneous nephrolithotomy channel.In order to enable doctors to accurately diagnose and treat kidney stones and reduce the risk of puncture,an automatic kidney segmentation algorithm framework based on the improved CV model is designed after comparing and analyzing a variety of conventional medical image segmentation algorithms,and then the three-dimensional images of kidneys,stones and related tissues are displayed using the three-dimensional reconstruction technology,which can be used to assist doctors to conduct comprehensive and accurate preoperative puncture planning and reduce the risk of surgery.The specific work is as follows:Firstly,the DICOM format CT data is read and bitmap format conversion is performed.After that,preprocessing such as denoising and window width and window level adjustment is also implemented.After image binarization,abdomen contour extraction and other steps,the detecting bed frame in CT images is removed.The canny algorithm,the regional seed growth and the morphological watershed algorithm based on seed point markers are respectively tried to segment the kidney.According to the comparison and evaluation of the segmentation results of each algorithm,it is finally established that the CV model based on the level set method is more suitable for the kidney segmentation of clinical CT images.Then,according to the existing problems in the traditional CV model algorithm,the regulation term is used to guarantee the stable evolution of the level set function.The curve evolution termination criterion is also added to reduce unnecessary iteration times and the accuracy of evolution is guaranteed at the same time.The improvement and optimization of the CV model algorithm are finally realized by adding the regulation term and the curve evolution termination criterion.Combining the gray distribution of the abdominal CT image and two phase segmentation of CV model with one level set function,the original image is windowed and adjusted to obtain a binary foreground image including kidneys,bones,and other tissues.According to the anatomical features of kidney organs and the contextual relevance of clinical CT images,the kidneys are automatically segmented and extracted frame-by-frame in the binary foreground image.The determination methods and manual correction interface for abnormal conditions are designed to ensure that segmentation results meet the actual clinical needs.Finally,on the basis of completing the segmentation of the kidneys,using the Otsu threshold method to segment and extract the stones within the kidney,effectively eliminates the interference of other tissues and organs,and the segmentation effect is good.By setting different colors and opacities using the VTK development toolkit,the three-dimensional reconstruction and display of kidneys,stones and abdominal skeleton are completed.The reconstruction results can intuitively and clearly show the shape of the kidneys,the size and shape of the stones,and their distribution inside the kidneys.After evaluation by multiple clinicians,it shows that segmentation and imaging results are accurate and can contribute to the preoperative path planning in the percutaneous nephrolithotomy.