一种用于电阻抗断层成像图像后处理的“像素化”自适应剖分方法

Pixelized adaptive mesh method for postprocessing of electrical impedance tomography images

目的:研究一种便于电阻抗断层成像(electrical impedance tomography,EIT)图像后处理的方法。方法:从算法角度入手,对现有剖分方法进行改造,选择相同直角三角单元作为基本的剖分单元,每2个直角三角单元组成一个方格,作为一个像素点。基于患者CT图提取颅脑轮廓,作为剖分边界。然后使用"像素化"剖分方法对提取出的轮廓进行自适应剖分,生成剖分模型后利用阻尼最小二乘算法完成重建,以此获取"像素化"的重建图像。结果:使用"像素化"自适应剖分方法,可以在不引入过大误差的情况下,获取与位图类似的"像素化"重建图像。"像素化"重建图像单元邻域关系明确,不用经过转换,可直接套用针对位图的图像处理算法。结论:"像素化"自适应剖分方法可以有效地简化EIT图像后处理过程,为电阻抗图像分析、指标提取奠定基础。

Objective To propose a pixelized adaptive mesh method for the postprocessing of electrical impedance tomography（EIT） images.Methods The algorithm of the existing mesh method was modified.The right triangle was used as the basic meth unit,and two triangles formed a square as a pixel.Craniocerebral contour was extracted from CT image to determine mesh boundary.Then adaptive mesh was carried out with pixelized method to generate a mesh model,and damped least square method was used for reconstruction to acquire pixelized reconstructed images.Results The proposed method resulted in pixelized reconstructed images similar to bitmap with no great errors introduced.The pixelized image units gained clear relationships between the neighboring ones,and image processing algorithms for the bitmap could be applied directly without any transformation.Conclusion The pixelized adaptive method simplifies the postprocessing of EIT images,and lays a foundation for EIT image analysis and index extraction.