基于形状插值的B超图像三维表面重建

3D Surface Reconstruction of B-scan Ultrasound Images Based on Shape-based Interpolation

B超图像由于受其成像机理的限制,其成像质量较差,无法采用与CT(计算机断层扫描)、MRI(磁共振成像)类似的三维重建方法。针对B超图像的缺点,采用基于形状的插值方法将其转换为等分辨率体数据。此外,对插值算法进行了适当改进,使重建的速度和效果得到明显改善。最后,分别采用移动立方体(marching cubes)算法、连接轮廓线、光线投射(ray casting)算法对B超图像进行三维重建。结果表明,移动立方体算法在重建的真实性和快速性方面都得到了更满意的效果。

Limited by the imaging mechanism, the quality of B-scan ultrasound images is lower than that of CT or MRI images, thus it is impossible to apply the similar 3D reconstruction methods that used in CT or MRI. Concerning with the limitation of B-scan ultrasound images, a shape-based interpolation method is used to change the volume data into an isotropic one. Moreover, the method is improved in this paper thus leads to faster speed and better effect. In addition, the volume data is reconstructed respectively by Marching Cubes, Connecting Contours and Ray Casting. It is shown that Marching Cubes can obtain more satisfactory fidelity and speed.