为了简化医院发票报销流程,提高财务管理效率,基于J2EE平台和改进ResNet模型构建医院智能化财务报销管理系统.系统在HTML、Java、JDBC、SQL Server 2019技术支持下,划分为3个技术层和5个功能模块.针对不同类型发票的报销问题,提出改进ResNet模型对发票进行分类识别,从而解决传统卷积神经网络存在过度拟合的缺陷.通过对比各种模型的使用效果,改进ResNet模型参数少,识别准确率更高,在训练时间和测试时间上也相对较少.对系统进行高并发访问情况下的性能检测,结果表明,系统在100人、点击率800次/s的高并发应用情况下,平均响应时间能保持在35 s左右,稳定性较好,可在实际工作中予以应用.
Design and Implementation of Hospital Intelligent Financial Reimbursement Management System Based on J2EE+ResNet
In order to simplify the hospital invoice reimbursement process and improve the financial management efficiency,an intelligent hospital financial reimbursement management system is built based on the J2EE platform and the improved ResNet model.The system is divided into three technical layers and five functional modules with the technical support of HTML,Java,JDBC and SQL Server 2019.Aiming at the reimbursement problem of different types of invoices,an improved ResNet model is proposed to classify and identify invoices,so as to solve the defect of over fitting in the traditional convolutional neural net-work.By comparing the use effects of various models,the improved ResNet model has fewer parameters,higher recognition accuracy,and less training time and testing time.The performance test of the system under the condition of high concurrent ac-cess shows that the average response time of the system can be maintained at about 35 s under the condition of high concurrent application with 100 people and 800 clicks/s,and the system has good stability,which can be applied in practical work.
intelligent financial reimbursement management systemJ2EEimproved ResNet modelclassification and identifi-cationaccuracy
万方数据
维普
