- 學位論文
利用影像處理技術配合流體力學模型建立來分析膽汁逆流
Using image processing techniques with fluid mechanics model to analyze the bile reflux
摘要
在本文中，肝膽道攝影影像被用於觀察預測膽汁流動的情形。由於膽汁是流動的液體，所以基於流體力學去建立膽囊到胃腸這一段的模型並區分成幾種膽汁逆流的類型或是沒有膽汁逆流的情形，然後利用假設的模型模擬出膽汁流動的情形。在動態影像分析中，光流技術被廣泛的利用於追蹤物體邊緣並且計算出物體的運動場。再加上光流法於一連串的2-D DISIDA影像以量化膽汁流動的情形。這個方法是利用兩張連續的影像來做運算，因此在每個像素，就可以得到兩張之間的速度向量並且可以清楚的知道整個結構的移動。根據每兩個影像得到的速度向量並且加上等高線法去描述流動的邊緣便可初步的了解膽汁逆流的流況。最後，比較光流法和的結果和流體力學的模型，並且希望可以找到評估此系統準確性和可靠性的標準。
並列摘要
In this context, the DISIDA CHOLESCINTIGRAPHY images are used to estimate the bile flow situation. Because bile is flowing fluid, so we based on the fluid mechanics to model that the stomach connects with the gall bladder in several typical of bile reflux status or no bile reflux, then simulate the bile flow situation by our assumption model. In dynamic image analysis, optical flow techniques are widely used to track object borders as well as evaluate the object motion field. After that we apply optical flow to the sequences of 2-D DISIDA images to quantify bile flow is presented. The method utilizes couples of consecutive frames to compute, for each pixel, the 2-D apparent velocity vector which characterizes the interframe motion. According to this vector with every couples of frames, and we use contour method to show the edge of flow. Therefore we can initially realize the afflux of bile reflux. Finally, we compare the optical flow result with the hydromechanics model, and hope to find the standard to estimate the degree of accuracy and reliability.