Research On The Application Of Medical Image Fusion Technique In Tumor Radiotherapy

Radiotherapy is a local treatment for malignant tumors with high energy radiation.The fundamental goal is to eliminate tumors as much as possible while protecting normal tissue and organ functions as much as possible,and strive to improve the long-term survival rate and quality of life of patients.With the development of science and technology,the requirement of precision of radiotherapy is higher and higher.All kinds of advanced medical imaging equipment are coming out,which plays an important role in the development of modern radiotherapy technology.Its application is permeated in various links such as radiotherapy scheme formulation,simulation positioning,radiotherapy planning and radiotherapy implementation.To provide clinical image data in various forms,to ensure the tumor in the course of radiotherapy clear and accurate.Due to different imaging methods and application scenarios,medical image data with different modes have a lot of redundant and complementary information.Therefore,how to effectively extract these complementary information and make full use of them is one of the major challenges facing us today.With the development of science and technology,image enhancement has become an important branch of image processing.Image fusion technology is the fusion of image data from different modes,through the use of redundant information to enhance reliability,through the use of complementary information to enhance clarity.The research of medical image fusion algorithm has made great progress in recent years,but there are still some problems worth exploring and researching.For example,how to design appropriate fusion rules to effectively remove noise interference;How to choose the right parameters to get the best effect.These problems need to be further solved.In this paper,the application of the fusion medical image to the delineation of radiotherapy target area as the research object,multi-modal medical image as the fusion image,from the perspective of guiding precision therapy,to explore the multi-modal medical image fusion algorithm and application.The full text consists of the following two parts:(1)A medical image convolution neural network fusion method is designed to realize multi-modal medical image fusion.In view of the widespread problems in medical image fusion,such as insufficient detail features and blurred edges between images of different tissues,this paper mainly discusses a medical image convolution neural network fusion method based on multi-scale feature residual network and attention mechanism.The network model consists of three units:feature extraction module,fusion module and reconstruction module.In the feature extraction stage,the structure of this method is improved by referring to the depth residual network,and the feature maps of different layers and scales are fused,which has a better feature extraction effect for targets of different scales.At the same time,the attention module is designed so that it can adapt to the output of feature maps of different scales.The expression ability of key channel features is enhanced,which makes important detail features in the image more prominent,and effectively solves the edge blurring phenomenon between different organizations.Finally,cross entropy loss combined with center loss was used to calculate,so that the distribution of sample features of each category in the sample space was more optimized,and the fusion accuracy of the model was further improved.The edge between different tissues in the fusion image is clear and rich in details.Experiments show that compared with the current advanced medical image fusion methods,the fusion effect of the proposed method is superior to that of the comparison method both from the subjective visual effect and from the objective evaluation index.(2)In order to verify the feasibility and effectiveness of the application of medical image fusion technology in tumor radiotherapy,the medical image after fusion was applied to delineate the tumor radiotherapy target area.Twenty patients with non-small cell lung cancer were selected in clinical trials to analyze the application of PET/CT fusion images in tumor radiotherapy target volume mapping,and observe and compare the difference of target volume(GTV)between all patients using CT mapping and PET/CT mapping.Differences in target volume(GTV)of different tissue types under CT and PET/CT mapping and changes in TNM staging of all patients after PET/CT imaging.The mean GTV volume of all patients sketched with PET/CT was(20.6±3.1)cm3,which was significantly lower than that of CT(25.1±2.2)cm3.There were significant differences in GTV volume between CT and PET/CT(P<0.05).Among all patients with NSCLC,PET/CT imaging indicated a change in stage in 8 patients(40%).The experimental results show that the use of medical image fusion technology can effectively distinguish and distinguish the tissue,so as to more accurately determine the target volume of radiotherapy for patients,improve the clinical efficacy.