With the advancement of cancer radiotherapy technology,the problems caused by organ movement and deformation during the treatment process become more and more prominent.In the course of radiotherapy,due to positioning errors,organ deformation and other reasons,the position and shape of the organ are changing during each treatment.In the case of high-dose gradients,even small organ deformations or positioning errors may cause the target area to be missed and endanger the organs to be involved in the high-dose area,resulting in uncontrolled tumors or excessive exposure of the endangered organs,resulting in patient health Great side effects.For the above situation,adaptive radiotherapy is usually used to solve.The key to adaptive radiotherapy technology is the selection of deformation registration algorithm.An effective deformation registration algorithm not only helps to derive the contour line of the target area,but also obtains a more accurate contour of the target area.It also helps the radiotherapy dose mapping.Based on this research,the LCC-Demons deformation registration algorithm and the B-spline deformation registration algorithm are used for the number of CT images in radiotherapy,and an improvement strategy is proposed for the shortcomings of the two algorithms to improve the registration speed and accuracy.The affine matrix and the deformation field obtained by the registration are used to complete the delineation of the doctor’s outline information,and then the accuracy of the registration algorithm is judged by comparing the DSC(overlap)and Hausdorff distance between the target area and the endangered organ before and after registration to make the registration quality Evaluation.The DSC overlap obtained by the deformation registration algorithm is greater than 0.8,which is a clinically acceptable range.For small anatomical tissues,organs and organs,the DSC overlap is between 0.7 and 0.84.The average Hausdorff distance obtained by the deformation registration algorithm is between 1 and 3 mm,and the surface deviation distanceis within its clinical error range. |