从图像中提取离散点状神经功能束边缘的研究

Study on Discrete Nerve Fascicular Groups Edge Extraction from Slice Image

针对神经切片图像中呈现离散点状分布的功能束,提出了一种提取其边缘轮廓的方法。首先采用聚类数目未知的动态聚类方法把功能束从原始切片图像中逐个分离出来,接着用改进的GVF-Snake模型逐个提取功能束的边缘,再把所得到的各个功能束边缘叠加到原始切片图像中。实验表明,这种方法不仅能有效处理功能束的分裂和汇合,而且边缘提取的收敛速度快,最终得到的功能束边缘光滑连续,满足神经三维可视化的要求。

To counter to the discrete pixels of nerve fascicular groups in a slice image,a new method was proposed to extract its edges.With the unsupervised dynamic clustering algorithm,the nerve fascicular groups were picked from original slice image one by one.Then the edges of the nerve fascicular groups were respectively extracted by the improved GVF-Snake model.Finally,all the edges of the nerve fascicular groups obtained above were added to the original image.The experiments show that the new method can process the split and merge of nerve functional fascicles very well and the improved GVF-Snake＇s convergence rate is faster than the traditional GVF-Snake.The final extracted edges of nerve functional fascicles are smooth and continuous and satisfy the requirement of nerve 3D visualization.