基于模式识别与灰度共生矩阵的神经元图像分类

Classification of Nerve Cells Based on Pattern Recognition and Features of Image Gray-level Co-occurrence Matrix

为了实现自动高效且结果准确的生物神经元识别,提出一种基于模式识别与图像灰度共生矩阵特征的神经元自动分类方法。该方法通过对生物神经元图像预处理,计算图像的灰度共生矩阵,统计各图像灰度共生矩阵属性值的平均值和标准差,构建生物神经元类别的特征空间,利用模式识别中的人工神经网络方法建立特征空间与神经元类别之间的映射关系。采用收集的160幅生物神经元图像对该方法进行实验分析,测试集的识别正确率达93.8%。研究结果表明,结合模式识别与图像灰度共生矩阵特征的生物神经元图像自动分类方法具有较高的准确性与可靠性。

An automatic neuron classification method based on pattern recognition and gray-level co-occurrence matrix features is proposed in order to realize the automatic,efficient and accurate recognition of biological neuron. In this method,the calculation of image gray-level co-occurrence matrix is based on the preprocessing of biological neuron images,the feature space of biological neuron category is established through the statistics of the average value and standard deviation of the gray-level co-occurrence matrix attribute values of the biological neuron images,and the mapping relation between feature space and neuron categories is obtained by means of artificial neural network method in pattern recognition. The automatic neuron classification method was tested using a test set of 160 biological neuron images,and the accuracy of this method is up to 93. 8%,which shows that the automatic classification method of biological neuron images based on pattern recognition and gray-level co-occurrence matrix features is accurate and reliable.