基于支持向量机的甲状腺结节能谱研究  

Energy Spectrum Scanning of Thyroid Nodules: A Study Based on Support Vector Machine

作　　者：杨创勃[1] 牛四强[2] 贾永军[1] 于勇[1] 段海峰[1] 贺太平[1] 

机构地区：[1]陕西中医学院附属医院CT室,陕西咸阳712000 [2]西安工程大学理学院,陕西西安710048

出　　处：《中国医学影像学杂志》2015年第3期231-234,240,共5页Chinese Journal of Medical Imaging

摘　　要：支持向量机(SVM)是一种基于统计学习中VC维理论结构和风险最小化理论的机器学习方法。本文对甲状腺结节临床病例的宝石能谱CT平扫数据进行分析,同时建立SVM的能谱平扫数据的诊断模型,并对实验指标进行约简,采用SVM预测分析。本文提出的诊断模型和实验方法能够为临床诊断甲状腺结节提供指导。PurposeSupport vector machine （SVM） is a machine learning method based on statistical learning theory of Vapnik-Chervonenkis （VC） dimension structure and risk minimization theory. We analyzed the gem spectrum CT scan data of patients with thyroid nodules and established the SVM diagnostic model. The experimental targets were then reduced and the forecast analysis was carried out based on SVM model. The diagnostic model and experimental methods were proved to provide guidance for clinical diagnosis of thyroid nodules.

关 键 词：甲状腺结节 体层摄影术 X线计算机 图像处理 计算机辅助 数据说明 统计 支持向量机 

分 类 号：R581.3[医药卫生—内分泌]