基于K-均值聚类的神经切片图像功能束类型识别研究

Research of clustering algorithm based on K-means for identification function bunch in neural biopsy dyeing image

针对周围神经物理切片图像中不同类型的神经功能束染色后显示出的特异性,提出了一种基于K-均值聚类的神经切片染色图像中神经功能束类型识别的方法。首先通过特征分析与提取决定以灰度均值μ和方差σ2作为纹理特征,然后运用本文算法对神经切片染色图像中的不同类型神经功能束进行聚类和识别。通过在人体周围神经组织切片图像上的实验证明,该算法能对神经切片染色图像中的神经功能束类型进行有效分类识别。

Aiming at the showing specificity in different types of nerve function after dyeing in peripheral nerve physical slice image, this article proposes a clustering algorithm based on K-means that is used for identifying the type of neural function bunch. Firstly, it decides average gray μ and variance σ2 as the texture characteristicsas through the analysis of characteristics and extraction, then clustering and recognition the different types of nerve function in neural slice dyeing images by using the algorithm of this aticle. The experiments in human peripheral nerve biopsy images prove this algorithm is of great use to for classifying and identifying the type of neural function bunch in neural biopsy dyeing image.