高效的白细胞分割算法

发布时间：2018-07-04 21:59

本文选题：目标检测 + 白细胞分割　；参考：《计算机工程与应用》2016年21期

【摘要】：白细胞分割是细胞形态学研究中一个重要、富有挑战性的课题。提出一种基于目标检测的白细胞分割算法。具体地讲,首先根据目标检测方法检测出白细胞,并由白细胞的位置信息得到包含白细胞的子图,然后运用多项式拟合的方法得到子图的灰度直方图的波谷值,再在子图上运用直方图阈值算法分割出细胞核。该算法在分割细胞核的过程中,既可以有效地避免血小板和红细胞等干扰,又能较容易地估计出阈值。对于白细胞的细胞质分割,将白细胞位置信息作为Grab Cut算法中人工交互部分,通过迭代法分割出白细胞的细胞质。实验结果表明,该算法能准确地定位白细胞,并根据白细胞的位置信息可以降低白细胞分割的难度,提高其分割的精度和分割效率。特别地,在Cellavision的白细胞图片数据库的实验结果表明,所提的白细胞分割算法对不同类别、不同染色剂和不同拍摄环境下得到的白细胞都能得到较好的分割效果,同时算法又还具有很好的泛化性。
[Abstract]:Leukocyte segmentation is an important and challenging subject in the study of cell morphology. A white blood cell segmentation algorithm based on target detection is proposed. Specifically, first, the white blood cells are detected according to the target detection method, and the subgraphs containing white blood cells are obtained from the location information of the white blood cells, and then the valley values of the gray histogram of the subgraphs are obtained by polynomial fitting method. Then the nucleus is segmented by histogram threshold algorithm on the subgraph. The algorithm can not only avoid the interference of platelet and red blood cell effectively, but also estimate the threshold value easily. For the cytoplasmic segmentation of white blood cells, the location information of white blood cells is regarded as the interactive part of Grab cut algorithm, and the cytoplasm of white blood cells is segmented by iterative method. The experimental results show that the algorithm can accurately locate white blood cells, and according to the location information of white blood cells, it can reduce the difficulty of white blood cell segmentation and improve its segmentation accuracy and efficiency. In particular, the experimental results in Cellavision's white blood cell image database show that the proposed white blood cell segmentation algorithm can achieve a better segmentation effect for white blood cells obtained from different categories, different staining agents and different shooting environments. At the same time, the algorithm also has a good generalization.
【作者单位】：中国计量学院理学院;嘉善加斯戴克医疗器械有限公司;
【基金】：国家自然科学基金(No.61272023,No.91330118) 浙江省自然科学基金(No.LY14A010027)
【分类号】：TP391.41;R329.2

【相似文献】