基于显微图像序列的细胞形态变化分析

Analysis of cellular morphological changes in image sequences

目的细胞的形态变化与细胞的生理特性密切相关,其定量的描述和分析对探究生命的生理或病理状态过程有重要意义。本文基于显微图像序列提取细胞形态的动态变化信息,以实现对细胞不同形态变化的定量描述及分类。方法采用运动历史图像和局部二值模式分别提取细胞轮廓和内部运动信息,并使用一系列不同尺度的时间窗口将上述特征映射为多时间尺度的特征矢量,再采用支持向量机对细胞的不同形态变化进行分类。通过对4组不同形态变化等级的小鼠淋巴细胞图像序列进行分类实验,以验证本方法的分类效果。结果对形态变化由缓慢到剧烈的4组淋巴细胞视频,分类精确度达到75%,能有效区分不同程度的细胞形态变化。结论与径向距离、Zernike矩、傅里叶描述子等常用的形状描述方法相比,本文方法更加全面地描述了细胞形态变化的动态信息,对细胞的多样性运动具有更好的适应性和稳定性。对细胞形态变化的分类,可用于异常细胞形态变化的检测,为疾病的早期诊断提供了客观依据。

Objective Cell morphological change is related to the cellular states and reflects the dynamic processes. So the quantitative analysis of cell morphology is critical for understanding the mechanism of physiological and pathological processes. This paper focuses on the feature extraction of cellular movements,aiming to achieve the quantitative analysis and classification of cellular morphological changes. Methods Motion history image is introduced to capture the deformation of cell boundary and local binary pattern describes the intracellular motions. Moreover,a series of temporal windows are adopted to encode the features into a multitemporal feature vector and combined with the support vector machine to classify the cells with variance deformations. Experiments are provided to assort the lymphocytes into four groups according to their morphological states,taken from the blood samples of mice. Results The classification accuracy achieves 75%and indicates the effective of the proposed framework. Conclusions The analysis of cellular morphological changes in images leads to a comprehensive characterization of morphological motion and is more suitable for heterogeneous motion representation. The proposed method is discriminative for variant morphological changes,which can be used for the detection of abnormal cell morphological changes and assist in early diagnosis.