Computer Vision Based on Fluid Equation-Reconstruction of 3D Flow Fields from Surface Flow image and X-ray Cineangiograms-

目标：使用流体方程作为约束，从表面图像和X射线Cineang图像重建流场。提出了一种从流场中获取符号描述的新模型和新算法。给出了层流和涡流的检测算法。通过我们的方法，我们从流场中获得了定性和定量的信息。提出了通过求解X射线投影图像序列中的流体凸起结构来估计流场的方法。将该方法应用于实际的DSA图像序列，对血流流场进行了估计。然后，我们提取具有临床意义的涡旋的位置或强度等特征值。提出了一种以表面图像序列和流体(Navier-Stokes)方程为约束的三维重建方法。对遮挡图像的三维血流进行了估计。我们应用于卡门涡旋的真实图像，并估计了遮挡图像数据的二维流。

Objectives : Reconstruction of flow fields from surface images and X-ray Cineang iograms, using fluid equations as a constraint1. We proposed a new model and new algorithm to aquire symbolic description from flow field. And we present our detecting algorithm for laminar flow and vorticies. Through our method, we obtain both qualitative and quantitative information from flow field.2. We presented techniques for estimating flow field by solving fluid salient structure from X-ray projection image sequence. Applying the methd to the real DSA image sequence, we estimated the blood flow field. Then, we extract feature values such as positions or strengthes of vortices expected as giving clinical signi ficance.3. We presented a new method of 3D reconstruction using surface image sequence and fluid (Navier-Stokes) equation as a constraint. We estimated the 3D flow for occlusive image. We apply to real images of Karman vortices and we estimate 2D flow of occlusive image data.