Active segmentation of 3D axonal images.
Active segmentation of 3D axonal images.
复制标题
3D 轴突图像的主动分割。
DOI：
10.1109/embc.2012.6346845
复制
发表时间：
2012
期刊：
影响因子：
--
通讯作者：
Ben-Yakar,Adela
中科院分区：
文献类型：
--
作者：
Muralidhar,GautamS;Gopinath,Ajay;Bovik,AlanC;Ben-Yakar,Adela
我们提出了一个主动轮廓框架分割神经元轴突的三维共聚焦显微镜数据。我们的工作是出于需要进行高通量实验，涉及微流体设备和飞秒激光，以研究神经再生和修复背后的遗传机制。虽然大多数主动轮廓的应用都集中在分割2D医学和自然图像中的闭合区域，但还没有很多应用集中在分割2D或更高维度的开放式曲线结构。我们在这里提出的活动轮廓框架将众所周知的2D活动轮廓模型[5]沿着投影成像几何学的物理原理联系在一起，以产生3D中的分割轴突。定性结果说明了我们的方法分割神经轴突的三维共聚焦显微镜数据的承诺。
We present an active contour framework for segmenting neuronal axons on 3D confocal microscopy data. Our work is motivated by the need to conduct high throughput experiments involving microfluidic devices and femtosecond lasers to study the genetic mechanisms behind nerve regeneration and repair. While most of the applications for active contours have focused on segmenting closed regions in 2D medical and natural images, there haven't been many applications that have focused on segmenting open-ended curvilinear structures in 2D or higher dimensions. The active contour framework we present here ties together a well known 2D active contour model [5] along with the physics of projection imaging geometry to yield a segmented axon in 3D. Qualitative results illustrate the promise of our approach for segmenting neruonal axons on 3D confocal microscopy data.