熱影像技術應用於頸動脈狹窄程度預測之研究

A Study of Using Thermal Image in Detecting the Level of Carotid Stenosis

林聖淵(Sheng-Yuan Lin)

指導教授 ： 蘇振隆

中原大學/工學院/醫學工程研究所/碩士(2004年)

https://doi.org/10.6840/cycu200400769

摘要

頸動脈狹窄是引起缺血性腦中風的重要危險因子，傳統評估方法不論是血管動脈攝影或是都卜勒超音波，核磁共振血管攝影，都有其侷限性以及耗時較長的缺點。證據顯示，同側臉部溫度的測量，可以反應同側血流量以及血流分布變化。本研究的目的是在利用紅外線攝影機拍攝人臉的熱影像，藉由影像處理和分析的技術來評估病人頸動脈的狹窄程度，作為腦中風病發的預測。
本實驗利用紅外線攝影機拍攝醫院腦血管住院病患的側臉熱影像，利用USB介面擷取灰階JPEG熱影像檔案到電腦儲存，並用Borland C++ Builder 6.0為開發工具建構分析影像的系統，選取臉部影像的臉頰、額頭、下顎等部位，計算其平均溫度和各部位溫度分布的標準差，探討其相關係數，以預測病患頸動脈的狹窄程度。
本研究分析依據ICA Flow超音波流速頻譜將25個病例分為阻塞低於50%和阻塞高於50%兩個族群。結果發現:阻塞大於50%的病例臉頰四周的部位平均溫度會比嘴唇以下的下顎部位之平均溫度明顯降低，而且臉頰部分的溫度分布標準差也會比較高；阻塞程度低於50%的病例臉部溫度分布則比較均勻，而且左右臉對稱性高。將下顎部位及臉頰部位的溫度差和根據NASCET標準計算出的頸動脈ECA阻塞程度作線性關係比較，發現ECA阻塞程度越高的病例，其臉部的溫度差距也越大。
藉由紅外線熱影像的檢查方法雖然還不能精確的決定阻塞程度超過50%的臨界值，但是相較於都卜勒超音波檢查方式時間明顯的縮短而且過程中不需接觸病人的身體。未來希望延續推廣此類方法，進一步增加病例，利用上述結果建立資料庫，可以有效作為快速評估頸動脈血流的篩檢性方法，並做為未來臨床事件追蹤的基礎。

關鍵字

臉部血流分布 ； 紅外線照影 ； 頸動脈狹窄 ； 影像處理 ； 都卜勒超音波

並列摘要

Carotid artery stenosis, as a major risk factor of ischemic stroke, is continuously lacking a rapid, effective screening method for the patients of stroke, high risk population, or even with early signs of stroke. All the traditional methods to assess carotid condition such as carotid angiography, duplex scanning, and magnetic resonance angiography, used to be time-consuming , are not adequate methods for screening. There is a lot of evidence in the literature that the ipsilateral temperature measurement could represent the facial blood flow and predict indirectly the carotid condition. This study tries to develop a system to analysis a facial thermal image and predict the level of carotid artery stenosis.
An infrared camera which produced by FLIR is used to photo the side facial thermal image, and then by using USB bus to transfer the gray JPEG files and store it. Borland C++ Builder 6.0 is used as the developing tool of system for image analysis. In this study, the specific region of face is selected and then the mean temperature and its standard deviation of face are calculated. In order to evaluate this developed system, we compare it with the results of duplex ultrasonic scanning.
In this study, we divide 25 cases into two groups, stenosis higher and lower than 50%, according to the duplex ICA flow velocity. The result shows that the average temperature of the area around cheek with is lower than the average temperature of mental proturberance for the high stenosis group. After compared the average temperature difference between cheek and mental proturberance to the stenosis degree of ECA carotid artery which is defined by the NASCET standard, positive linear correlation was found.
The system can’t define the threshold of index which can exactly predict the stenosis of 50%. However, the response of system is faster than duplex scanning and don’t need to touch the patient’s body. In the future, we hope to continue such study and compare the result with the traditional methods to assess carotid condition. Besides, we wish to increase cases to develop a data base and an effective screening method for all the patients of stroke, high risk population, or even with early signs of stroke as well as clinical follow-up.

並列關鍵字

image processing ； carotid stenosis ； facial blood flow ； infrared image ； Duplex scanning

參考文獻

[11] 陳加興,龔顯琦,陳啟昌,李三剛,謝彩堂,”用電腦斷層血管攝影曲線多平面重組法和軸位切面影像評估頸動脈狹窄”,中華放射醫誌　24(6)：239-242,1999.

[3] JF Head and RL Elliott ,”Infrared imaging: making progress in fulfilling its medical promise”,IEEE Engineering in Medicine and Biology, 21(6):80-85,2002.

[5] M Gautherie and CM Gros,”Breast thermography and cancer risk prediction”, Cancer, 45:51-56,1990.

[7] L Pavlidis and J Levine,"Thermal image analysis for polygraph testing", IEEE Engineering in Medicine and Biology,21(6):56-64,2002.

[10] European Carotid Surgery Trialists' Collaborative Group, MRC European Carotid Surgery Trial: Interim Rsults for Smptomatic Ptients With Severe or Mild Carotid Stenosis, Lancet, 337:1235-1243,1991.