摘要
血清(血浆)是临床化学分析中最重要和最为常见的分析测试标本。血清(血浆)的化学组成与人体组织器官(特别是肝脏)和血液中各种细胞(如淋巴细胞、粒细胞、单核细胞等)密切相关,与整个机体的生理、病理状态密切相关,因此蕴藏着与疾病的发生与发展、临床诊断及治疗、预后评估等等大量的有用信息。试验将一定的化学试剂与正常和肝病(肝炎、肝癌)患者血清作用,生成"化学试剂-血清结晶",然后对结晶图像进行曲线波变换(Curvelet transformation),完成结晶图像的特征提取,用支持向量机(Support Vector Machine)对结晶图像进行分类识别。初步实验结果表明,正常与肝炎患者的化学试剂,血清结晶图像分类识别率可达81.65%,正常与肝癌患者的化学试剂-血清结晶图像分类识别率可达85.47%,肝炎患者与肝癌患者的化学试剂,血清结晶图像分类识别率可达79.12%。实验应用图像多尺度几何分析中的曲线波变换及支持向量机对正常和肝病患者血清"化学试剂,血清结晶"图像进行了较好的分类识别,不需要昂贵的仪器和试剂,是一种相对简易的方法,同时拓展了曲线波变换的应用范围。进一步研究提高识别率,期望在肝脏有关疾病的诊断、治疗及预后判断以及在人体健康状态辨识等方面发挥一定的作用。
Serum（plasma） is the most important and the most common clinical chemical analysis test specimen.The chemical composition of serum （plasma） is closely related to human tissues,organs（especially the liver） and all kinds of cells in blood,such as lymphocytes,granulocyte,monocytes, etc.It is also closely related to the condition of physiology,pathology.Thus it contains a lot of useful information—the occurrence and development of disease,clinical diagnosis and treatment,prognosis assessment and so on.We mixed certain chemical reagent with the serum from normal and liver disease（hepatitis and liver） patients,and generating＂chemical reagent-serum crystallization.＂Then curvelet transformation was done,as well as characteristic was extracted,and the pattern recognition was made with Support Vector Machine.Results show that the recognition rate is close to 81.65%between normal and hepatitis patients chemicals-serum crystallization,85.47%between normal and liver cancer patients chemical reagentserum crystallization,and 79.12%between hepatitis patients with liver cancer patients chemicals-serum crystallization.Curvelet transformation and support vector machine are good classification identification method used in serum＂chemical reagent-serum crystallization＂between healthy people and patients with liver diseases.This method is easy to carry out without expensive equipment and reagent,and expands the application of curvelet transformation.The improvement of recognition rate is expected to be beneficial to the diagnosis,treatment,and prognosis assessment of liver diseases,as well as to the recognition of health status of human beings.
出处
《计算机与应用化学》
CAS
CSCD
北大核心
2011年第4期447-450,共4页
Computers and Applied Chemistry
基金
浙江省科技厅(200914024)
