点击次数：

所属单位：自动化学院

发表刊物：INFRARED PHYSICS & TECHNOLOGY

关键字：Osteoarthritis Chemometics Support vector machine-discriminant analysis (SVM-DA) FTIRI

摘要：Fourier transform infrared spectroscopic imaging (FTIRI) technique can be used to obtain rich information about the chemical components in biological tissue such as articular cartilage, by employing quantitative chemometrics methods. In this study, the imaging technique was used for spectroscopic imaging on cartilage sections. The support vector machine discriminant analysis (SVM-DA) was then employed in imaging data analysis to distinguish among healthy and osteoarthritic (OA) articular cartilages at different degeneration stages for the first time. Briefly, the infrared spectra were extracted from the FTIR images and imported into Unscrambler software for the SVM-DA model construction and prediction. When the 3rd-polynomial kernel function was chosen as the kernel function and the parameters C in each range of 2782.5-10000, 2.3714-10 and 0.0032-0.01, respectively, and the corresponding optimal G was equal to 0.1, 1 and 10, the healthy, OA-8W (8 week after surgery) and OA-2Y (2 years after surgery) cartilage samples were effectively differentiated by the SVM-DA method with high accuracy of 100% for the training set and 86.67% for the prediction group. FTIRI with the use of chemometrics (SVM-DA) may become an effective method to distinguish multi-stage tissue degradation in the future.

ISSN号：1350-4495

是否译文：否

发表时间：2018-08-01

合写作者：Zhai, Ming-Yang,Mao, Zhi-Hua,Lu, Yan-Fei,尹建华

通讯作者：王潇,尹建华