一种基于H直方图变换的白细胞图像分割方法

Leukocyte Image Segmentation Algorithm Based on Hue Component Histogram Transformation

对检出的大量样本细胞图像进行统计分析,发现白细胞在Wright染色后着色效果不一致时色彩分量差异的规律。通过建立数学表达模型,提出了一种基于H直方图变换的白细胞图像自动分割算法。对细胞图像的色度分量直方图进行适当变换,有效削弱了血液染色的不一致性对白细胞分割的影响,实现了白细胞胞浆区域的良好分割,实验结果验证了所提分割方法的可行性和实用性。

With analyzing large number of leukocyte sample images statistically,some features of color component differences caused by Wright stain inconsistent effect are found.By establishing its mathematical model,a leukocyte image automatic segmentation algorithm is proposed.By properly transforming the histogram of hue component in leukocyte images,the infection for leukocyte segmentation caused by staining inconsistency has been weakened obviously and the leukocyte segmentation accuracy has been improved.The experimental results verify the feasibility and practicability of the proposed segmentation method.