請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/89658
標題: | 以近紅外光譜及類神經網路定量內頸靜脈血氧飽和度變化量 Quantification of Changes in Internal Jugular Vein Oxygen Saturation by Near Infrared Spectroscopy and Artificial Neural Network |
作者: | 謝昕原 Hsin-Yuan Hsieh |
指導教授: | 宋孔彬 Kung-Bin Sung |
關鍵字: | 漫反射光譜,內頸靜脈,血氧飽和度,蒙地卡羅演算法,擾動式蒙地卡羅,類神經網路, Diffuse Reflectance Spectrum,Internal Jugular Vein,Oxygen Saturation,Monte Carlo Algorithm,Perturbed Monte Carlo,Artificial neural network, |
出版年 : | 2022 |
學位: | 碩士 |
摘要: | 本研究主要目標在於以非侵入式的近紅外光量測受試者頸部組織的漫反射光譜,並透過蒙地卡羅演算法及類神經網路的方式來建立預測模型,用以定量內頸靜脈的血氧飽和度的變化量。
量測系統上,使用LED光源,選擇5個分析波長,分別為730、760、780、810、850 (nm),搭配自製的偵測光纖束來對人體進行量測。蒙地卡羅的組織模型則利用超音波影像來輔助建立,使其能夠更接近的真實的組織結構,而獲得更加準確的模擬資料。訓練神經網路所需要的大量模擬光譜則藉由白蒙地卡羅及擾動式蒙地卡羅來進行加速。建立一套能夠分析並定量內頸靜脈血氧飽和度變化量的預測模型。 本研究所建立之模型可預測出內頸靜脈血氧飽和度變化量,誤差小於4%,但預測模型會受到其他淺層組織的影響,此問題可用增加較短通道的偵測光纖來改善。活體實驗上,以過度換氣的方式來調變受試者的血氧飽和度,並以本研究建立之預測模型分析,其結果與預期吻合,可說明本研究之方法及組織模型是能夠用來預測活體組織的重要生理參數:內頸靜脈血氧飽和度變化量。 The goal of this study is to measure the diffuse reflectance spectrum (DRS) of the neck tissue of the subject non-invasively, with the Monte Carlo algorithm and neural network to quantify the Changes in blood oxygen saturation in the internal jugular vein (IJV). On the measurement system, an LED light source with a wavelength range of 550 ~ 1100 (nm) and a center wavelength of 700 nm is used, and a self-made detection fiber bundle is used to measure the human body. Ultrasound imaging is used to aid in the creation of tissue models of Monte Carlo, so that it can be closer to the real tissue structure and obtain more accurate simulation data. The large number of simulated spectra needed to train the neural network is accelerated by white Monte Carlo and perturbed Monte Carlo. To establish a predictive model capable of analyzing and quantifying changes in internal jugular vein oxygen saturation. The model established in this study can predict the changes in the oxygen saturation of the internal jugular vein with an error of less than 4%, but the prediction model will be affected by other superficial tissues. This problem can be improved by adding a detection fiber with a shorter channel. In the in vivo experiment, hyperventilation was used to modulate the blood oxygen saturation of the subjects, and the prediction model was used to analyze the results. to predict unknown physiological parameters (ΔSijvO2) of living tissue. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/89658 |
DOI: | 10.6342/NTU202204054 |
全文授權: | 同意授權(全球公開) |
顯示於系所單位: | 生醫電子與資訊學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-110-2.pdf 此日期後於網路公開 2025-09-30 | 4.32 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。