基于三维梯度幅值的CT图像体绘制

Volume Rendering for CT Images Based on 3D Gradient Amplitude

为了在CT影像体绘制过程中保持感兴趣区域三维空间的局部形状特征,提出了一种基于三维梯度幅值的传递函数设计方法:利用边缘提取和去噪分割等处理得到感兴趣区域的完整边缘结构,计算体数据中图像序列像素点到边缘处的距离,以此距离值纳入Kindlmann的传递函数生成框架,得到新的传递。与Kindlmann提出的半自动传递函数生成算法进行体绘制对比实验,结果表明:原始方法对图像噪声敏感等问题得到明显改善,能有效突显出感兴趣的区域。

In order to maintain the partial shape characteristics of the three-dimensional space of interest in the process of CT images volume rendering, a transfer function method based on the three dimensional gradient amplitude is proposed. Edge detection and denoising segmentation are employed to obtain the complete edge structure of the interested area. The distance between the image sequences pixels and the edge in the body data is calculated, so that the framework can be formed with the distance data brought into the Kindlmann transfer function. Therefore, a new transfer function is achieved. The comparison experiment with semi-automatic transfer function proposed by Kindlmann shows that problems like image noise sensitivity can be obviously solved by original methods, which can also highlight the interested parts.