Design And Implementation Of Lung Cancer Assisted Diagnosis Based On Artificial Intelligence

According to the investigation and study,the number of lung cancers found in medical examinations has increased year by year in recent years,but the number of people who choose to be hospitalized is relatively small.About 75%of patients are already in the middle and late stages of lung cancer when they are diagnosed,and their chances of curing are greatly reduced.Therefore,early diagnosis of lung cancer is of great significance to reduce the mortality caused by lung cancer.On the other hand,in the face of a large number of patients,clinicians have a considerable workload,and need to screen patients with lung cancer first,and then treat patients according to their own experience.However,in China,it takes a long period of time to train an experienced clinician.In summary,in the medical field,research on the use of artificial intelligence technology for medical data analysis and assisting clinicians in screening diagnosis and treatment of lung cancer has important theoretical significance and practical application valueAfter analyzing the current deep learning methods,this article finds that researchers at home and abroad mainly make lung cancer predictions for CT images of the lungs,ignoring the CT image descriptions and test reports given by radiologists,especially the test reports,so that some information will be lost.Considering the above problems,this paper designs a novel text and image multimodal learning lung cancer-assisted diagnosis scheme.This scheme is different from the existing methods.It is based on the CT image,the CT image description given by the radiologist,and the inspection report for multimodal fusion.The main realization point is to pre-process the image part first,and then use Resnet network modeling;CT image description part uses natural language processing technology for word segmentation,pre-training,and modeling;the inspection report is modeled by multi-layer perceptron;After three parts merge Experiments verify that the accuracy of multimodal methods based on text and CT images is 3%higher than that of single-modal methods based on CT images.This shows that CT images are still the main information for the diagnosis of lung cancer.The test results are added to the model as supplementary information,which can greatly improve the accuracy of the modelBased on the multi-modal learning method of text and images designed in this paper,the lung cancer assisted diagnosis system is designed and implemented.Fully integrate multi-modality with computer-aided diagnosis to help clinicians screen and diagnose patients.The system can quickly realize data preprocessing,auxiliary diagnosis and judgment of lung cancer,diagnosis information entry,and query of patients' past history and cases.The implementation of the system reduces the workload of clinicians and improves their work efficiency.It also provides clinicians with a comprehensive observation and diagnosis of patients.At the same time,lung cancer patients can also understand their own situation in time.