Study On Needle Recognition Technology In Brachytherapy

As a low-risk surgery with high efficiency in killing tumors,brachytherapy has received increasing attention.With the development of social science and technology,treatment planning systems are becoming increasingly active in medical technology,especially in areas of lesions that are difficult to observe in the human body during brachytherapy surgery.Treatment planning systems can be used to simulate preoperative planning to achieve accurate treatment goals.This article focuses on the mismatch between preoperative and intraoperative plans in the surgical process of lung brachytherapy particle implantation,as well as the issue of intraoperative needle deviation.The main research contents include: research on preoperative and postoperative Chinese medicine image registration technology,performance comparison of mainstream linear segment extraction algorithms on medical images and research on intraoperative needle path recognition technology.In order to solve the problem of intraoperative failure of preoperative plans in current domestic treatment planning systems,this paper proposes a multi-resolution strategy based on preoperative and surgical Chinese medicine image registration technology,which uses wavelet transform to construct an image pyramid while improving computational efficiency,and matching it with a perfect optimization strategy.To further enhance the accuracy of image registration,a simplex algorithm is used to register the high-resolution layers of the image pyramid model accurately,and a complete registration process is constructed.Finally,the accuracy and efficiency of the algorithm are verified through medical image registration experiments.Aiming at intraoperative needle path recognition technology algorithms,the current mainstream linear segment extraction algorithms are studied.The primary purpose of this study is to compare the effectiveness of the EDLines algorithm and LSD algorithm in extracting linear segments from medical images.In past research and development,visual judgment has been used to compare the effectiveness of linear segment extraction,lacking an objective and reasonable evaluation system.This paper objectively evaluates the comprehensive performance of these two algorithms by using visual judgment,first interpreting and then verifying with a straight-line matching algorithm.It also lays a foundation for the accuracy and effectiveness of subsequent needle path recognition algorithm experiments.In order to solve the problem of needle path deviation caused by factors such as intraoperative tumor swelling and bleeding,and to assist doctors in obtaining the actual position of the implanted needle,an intraoperative needle path recognition technique was proposed.This algorithm improves on the traditional EDLines algorithm,replacing the original least square method with a random sampling consistency algorithm,improving the accuracy and efficiency of the algorithm.The Helmholtz principle is used to eliminate false linear segments,and the position of the needle tip is determined by searching for the gray level changes around it.Finally,it was verified through actual cases that over 97% of needle axis angle deviations are less than 1.2 °,and all needle tip position errors are less than 0.8mm,fully meeting the accuracy requirements of clinical surgery.