摘要
基于CT图像的心脏三维重建在医学影像分析领域需求大,但难度高。重构的3D心脏虚拟模型可有效支持医学研究以及临床决策。但由于噪声影响,单使用分割算法往往不能得到理想重建效果。提出了图像预处理和分割算法相结合,采用基于总变差变分去噪模型预处理CT心脏图像,通过OTSU阈值方法分割预处理的图像。以VTK工具包为基础,使用Marching cube(MC)算法和光线投影法对处理后的图像进行三维重建,得到CT心脏图像的三维虚拟模型。实验结果表明:图像预处理效果直接影响分割的效果,三维重建是实现医学图像的三维可视化的关键。
Three-dimensional (3D) reconstruction on medical CT heart images is a demanding but tough area in medical image analysis. The reconstructed 3D model can be very useful to support medical research or clinical decision marking. However, due to the inherent noise existing in CT images, it is hard to obtain satisfactory reconstruction results by using segmentation methods alone. In this paper, we propose to combine the preproeessing and segmentation methods. Specifically, the classical total variation model is first employed to remove noise from images. These noise free images are then segmented via the efficient OTSU thresholding method. Based on the visualization toolkit (VTK), the marching cube and ray casting algorithms are applied to the processed images so as to generate the 3D virtual model. Numerical experimental results validate the effectiveness and performance of the methods on the real medical CT heart images.
出处
《重庆理工大学学报(自然科学)》
CAS
2016年第12期102-107,共6页
Journal of Chongqing University of Technology:Natural Science
基金
黑龙江省留学归国基金资助项目"个性化web图像检索技术研究"(LC2012C06)
关键词
医学图像
总变差
OTSU
三维重建
三维可视化
medical images
total variation
OTSU thresholding
3D reconstruction
3D visualization