基于综合灰度共生矩阵的显微细胞图像纹理研究

Research on Micro-cell Image Based on Synthesized Gray Level Co-occurrence Matrix

灰度共生矩阵法是图像纹理分析中一种十分重要的方法。由于受到方向性和细胞本身特性的影响 ,传统的灰度共生矩阵法不能完整描述显微细胞图像的纹理 ,运用综合灰度共生矩阵可以有效地解决这个问题。根据显微细胞图像特点 ,计算出 6个由综合灰度共生矩阵导出的纹理特征参数 ,并对它们进行归一化处理。通过分析这些归一化纹理特征值 。

Gray level co-occurrence matrix is a very important method in image te xture analysis. Influenced by features of directions and cells, traditional algo rithm of gray level co-occurrence matrix can't perfectly describe the texture of micro-cell image. Synthesized gray level co-occurrence matrix can be utilized t o solve this problem. Due to the characteristics of cell image, 6 textural featu res produced by synthesized gray level co-occurrence matrix are calculated, then normalized. By analyzing the normalized textural features, different kinds of c ells in cell image can be classified and described.