基于自适应滤波的淋巴细胞图像分割算法研究

Research on Image Segmentation Method in Marrow Lymphocyte Cells Based on Self-Adaptive Mathematical Morphology Filter and Watershed Algorithm

研究细胞图像分割精确度问题。针对现有医学细胞图像分割算法普遍存在噪声敏感、难以处理粘连细胞以及时间复杂度高等问题,提出基于自适应形态滤波增强的淋巴细胞图像分割算法AMWS。首先采用自适应形态滤波增强算法对淋巴细胞图像进行去噪增强,既去除噪声又保护细胞图像的细节,然后采用双阈值分割对细胞进行粗分割,对重叠细胞进行标记分水岭细分割,完成整个分割过程。仿真结果表明,AMWS算法能够较好地完成淋巴细胞图像的分割,有效避免了现有图像分割方法的缺陷,弥补了现有细胞分割算法的不足,分割效果优于目前同类分割算法。

How to improve the quality of medical images segmentation is an important in researchs. With the rap- id developments of medical imaging technology, using computer technology to process medical images to alleviate medical staff workload and improve the diagnostic quality has become a hot international research direction. But there are lots of drawbacks in current methods, such as over segmentation, cell overlap, noising and so on. These short- comings lead to low quality of medical images segmentation. A smart AMWS algorithm was proposed based on self- adaptive mathematical morphology filter and watershed theory. The mathematical morphology processing was adopted to enhance the image, then serf - adaptive bi - threshold segmenting can be used to discrete cells, and watershed segmenting can be used to overlop cells. Simulation experiments show that this novel AMWS algorithm has good seg- mentation effect in not only discrete cells but also overlop cells than other image segment methods in marrow lympho- cyte cells.