The Research Of Calculus Target Positioning And Tracking Algorithm Based On CT Images

CT images are widely used in field of medical image-guided therapy, and inwhich play a huge role. Currently, the main method for kidney stone tracking andpositioning is based on the CT images, then the approach and analysis are used onthem. Due to the influence of respiration motion, the location of the kidney stone hasbeen changing frequently. Meanwhile, due to the system latency (such as imageacquisition time, data transfer time, and time that the machine takes to process thedata), and the fact that if the patients are exposed to the X-ray, there may cause greatunnecessary harm to their bodies, which also does harm to the other organs, so lessX-ray pictures should be guaranteed in the clinical application. The traditional processof kidney stone lithotripsy is primarily based on the doctors’ subjective experience totrack the stone and locate the stone. Sometimes, more time may be needed in thetherapy. While, the new method for the therapy is developed through good positioningand predicting algorithm to compensate the errors which are caused by the systemlatency. Owing to this, the accuracy of the stone location can be improved and thecurative effect is good.This article uses the coordinate conversion to locate the kidney stone targets’position. This method is much easier than other methods, and it has small amount ofcalculation. Meanwhile, the spatial coordinate system is established based on theultrasonic lithotripsy machine which is actual used in the treatment, and this can makethe process simple. Moreover, during the process of the collecting the projected images,the geometric distortion correction can ensure the result’s reliability.This article studies the mean position tracking algorithm, and an adaptive dataprocessing prediction algorithm is proposed, which is fully automatic, has a strongrobustness, and can guarantee the prediction accuracy. By using the ellipse fitting fordata processing, the effects of the noise can be reduced and outliers and missing values can be handled better. Moreover, on the basis of ellipse fitting, the sliding windowfunction, discounting factor and weights are used to achieve the adaptive algorithm.Eventually, owing to the connection of the sliding window function, discounting factorand weights, the historical data can be well used to help predict the target position.A series of kidney CT images acquired from different respiration period are testedby using the proposed algorithm, and the quantitative and qualitative evaluationsdemonstrate that our method can locate the kidney stone target position precisely andwell predict the future position. Meanwhile, the proposed method can tolerate thewrong data which is generated during the process, and the error analysis proves thebetter adaptability of this method.