基于主凹点检测的血细胞图像去粘连分割算法研究

Cell cluster splitting in blood cell images based on main concave point detection

针对医学中血细胞图像研究中粘连细胞难以分割的问题,提出一种基于主凹点检测的分割算法。通过滤波预处理去除图像的噪声以改善图像质量,基于改进的活动轮廓模型初步提取细胞轮廓,通过寻找主凹点的方法准确定位粘连细胞凹点位置,标记并融合细胞图像轮廓、粘连形状等特性,实现粘连细胞分离。实验结果表明,该方法具有很好的分割准确度和完整度,且该算法具有普适性。

Aiming at the problem of cell cluster splitting in the research of blood cells in medical image,it puts forward a segmentation algorithm based on the concave point detection. Firstly,through filtering pre-processing to remove the noise of the image and to improve the image quality. Then,extracting the contour in the early based on improved active contour model. Finally,positioning the cell adhesion concave point accurately by using the method of looking for the main concave points. Experimental results show that this method has good accuracy and integrity of segmentation,and the algorithm is universal.