- 學位論文
動態顯微影像估測微循環血流速
Microcirculation Flow Rate Measurement Based on Microscopic Images
摘要
微循環研究主要是瞭解微循環之生理機能，而透過局部觀察組織中微血管的血流速作為數據的來源。為了研究局部微血管中的流血速，故提出了動態影像估測的方式作為血流速量測的方法，文中量測了螢光老鼠大腦微循環影像、真實小鼠皮膚微循環影像和青蛙蛙蹼微循環影像，影像的擷取為非侵入式的攝影方式，利用攝影顯微鏡以影像取樣率每秒30張影像和空間解析度為1.42µm所拍攝下來，以獲得微循環血流速之動態影像資訊。 傳統的估測方式有兩種，光流法與相關係數法，光流法只能處理變動量小的流速，相關係數法只能處理細血管的影像來源。本論文方法的提出，主要是能夠處理每張影像之間變化量大，且能不用考慮血管的寬度，就能準確的估測到血管中的流速，使用了改良式的區塊比對法與光流法的結合來解決動態影像分析上的問題。 微循環血流速估測的結果中，對於螢光老鼠大腦影像血流速的量測，其血流速的量測結果介於204~444 µm/sec，青蛙蛙蹼影像流速的血流速量測介於54~137 µm/sec，老鼠皮膚影像血流速量測為38~254 µm/sec，從三組不同影像的結果中與文獻中所記載速率的比較上，估測出的流速都在真實流速的範圍內，證實本方法確實能準確估測微血管中的血流速，在微循環的研究將可提供此方面的生理數據做為參考與評估。
並列摘要
Microcirculation flow rate can respond the physiological function immediately. In order to study the local microvascular flow rate, we need to catch the dynamic data by microscope. In our research the frame rate is 30 frames/sec and the image resolution is 1.42μm/pixel, and there are three different sources, the rat brain microcirculation fluorescence image, the mouse skin microcirculation images and the frog webbed microcirculation imaging. There are two traditional ways to estimate the flow rate. The first one is optical flow that can only estimate the small movement. The other one is cross correlation that can be only applied on vessels with diameter of one red blood cell. In our research, the way is combined with Block-Match, TPS and optical flow. We can solve above two situations in dynamic images. In microcirculation blood flow rate estimation results, the flow rate measurement of the rat brain fluorescence image is between 204 ~ 444μm /sec, the flow rate measurement of the frog webbed images is between 54 ~137μm / sec, and the flow rate measurement of the mouse skin image is between 38 ~ 254μm / sec. The results compared with the real rate described in the literature are matched. It is confirmed that the present method can accurately estimate the blood flow rate in the microvessels.