传感技术学报2019，Vol.32Issue(1)：29-35,7.DOI:10.3969/j.issn.1004-1699.2019.01.006
基于模板匹配的肺癌呼出气体标志物识别研究
Study on Recognition of Lung Cancer Exhaled Gas Markers Based on Template Matching
摘要
Abstract
The color characteristic chromatogram formed by lung cancer exhaled gas markers is the an important medium for the application of porphyrin sensor arrays in the early screening for lung cancer. Therefore, the reasonably designed pattern recognition method based on characteristic chromatogram is the key to the realization of correct recognition of lung cancer markers. Aiming at the response characteristics of the characteristic chromatogram, a template matching recognition method is proposed, which does not only consider the number of response and color difference of the sensor units in characteristic chromatogram, but also assigns different weights to the sensitive response units and the non-sensitive response units in the calculation of the matching degree. Recognition tests were respectively carried out to samples of six types of lung cancer exhaled gas markers, samples of exhaled gas from 20 cases of lung cancer patients and 16 healthy people. The experimental results of samples of six types of lung cancer exhaled gas markers show that the method has a recognition accuracy of up to 94.17％, and compared with the common cluster analysis method, its detection limit is as low as 150×10-9. The experimental results of samples of actually exhaled gas from 36 people show that there are significant differences in the characteristic chromatogram of exhaled gas between lung cancer patients and healthy people. At the same time, the template matching method can be used to better distinguish the lung cancer patients from healthy people. This study has important application value for detection method for lung cancer exhaled gas based on porphyrin sensor array.关键词
卟啉传感器阵列/模式识别/模板匹配/肺癌标志物/特征图谱Key words
porphyrin sensor array/pattern recognition/template matching/lung cancer markers/characteristic chromatogram分类
信息技术与安全科学引用本文复制引用
易鑫,罗小刚,张承丹,侯长军,霍丹群..基于模板匹配的肺癌呼出气体标志物识别研究[J].传感技术学报,2019,32(1):29-35,7.基金项目
国家自然科学基金项目(81271930,No.81171414) （81271930,No.81171414）
国家科技支撑计划项目(2012BAI19B03) （2012BAI19B03）
