为了更好地完成骨髓细胞图像的分类识别,在对临床血液学及细胞学图谱中的细胞图像进行切分的基础上,提出了一种基于分形维数和图像重心的算法,用来提取每一细胞图像的分维特征.由于图像的纹理不仅表现在图像结构上具有某种统计意义的相似性,还反映在彩色分布方面,因此,可从骨髓细胞的真彩色图像中提取一个新的彩色识别分量图像C,并与HIS彩色空间的饱和度分量图像S一起进行分维特征提取.实验数据表明,由于不同类型骨髓细胞图像的分形参数具有不同程度的差异,因此分形参数对于某些类别的骨髓细胞具有较好的区分能力.

This paper proposes a new recognition algorithm based on fractal dimension and image center of gravity after the segmentation of marrow cell images in the clinical haematology and cytology atlas which can be used to extract. The fractal feature of every marrow cell image. Since the texture of a color image contains not only certain statistical similarity but also the color distributions, a new color parameter is extracted from the color images of marrow cells and discussed with parameter S of HIS for the fractal dimension. The experimental result shows that different marrow cell image has the certain discrimination in fractal dimension and some images can be recognized for classification.