摘要
针对目前新生儿胆红素含量升高造成的新生儿各种病症的弊端,考虑到传统新生儿胆红素水平测试所带来的患者创伤与操作不便捷的问题,在VGGNet-16神经网络的基础上,提出了一种改进的VGGNet-16网络算法,对已有的新生儿胆红素医疗记录数据进行分类研究,不同的类别具备一定范围的胆红素水平值;经过实际验证,该算法在克服传统新生儿胆红素水平测定不便捷的同时,保证了新生儿胆红素水平测试的高准确性,平均准确率达95.56%,为新生儿胆红素水平测量方法提供了思路,具有较高的推广应用价值。
In view of the shortcomings of various neonatal bilirubin levels caused by neonatal bilirubin,taking into account the problems of patient trauma and inconvenient operation caused by traditional neonatal bilirubin level test,an improved VGGNet-16 network algorithm was proposed to classify the existing neonatal bilirubin medical record data.Different categories have a certain range of bilirubin level values.After actual verification,the algorithm overcomes the inconvenience of traditional neonatal bilirubin level determination,and ensures the high accuracy of neonatal bilirubin level test.The average accuracy rate is 95.56%.It provides a way of thinking for the measurement method of neonatal bilirubin level.High promotion and application value.
作者
刘国玉
王东颖
候桂军
Liu Guoyu;Wang Dongying;Hou Guijun(Tangshan Maternal and Child Health Hospital,Tangshan 063000,China)
出处
《计算机测量与控制》
2020年第7期55-58,90,共5页
Computer Measurement &Control
基金
河北省卫生厅科研基金项目(20181331)。
