小波多尺度分割算法在细胞图像上的应用

摘要

In this paper, multi-scale edge detection based on wavelet transform was studied. Before the traditional edge detection algorithm was improved, we adopted an adaptive smooth filtering method. According to cellular image structure characteristics and clinical medicine morphology parameters measure accuracy, we proposed a method based on third-order B-spline Dual wavelet multi-scale edge detection. First, we detected the edge of typical sparse cell with B-gpline wavelet transform edge detection, then we used image morphology and threshold segmentation to segment the interested cells. Finally, we used morphology method measured the segmented cell, and obtained cell' s morphological parameters. We compared the actual value with the measured value by the presented method. The results show that the absolute error is small, and the measured value satisfies the requirements of clinical medicine.%关于数字图像优化分割问题,由于医学病理研究需要准确测量细胞面积大小.针对细胞图像结构特点,以及临床医学形态参数测量精度要求,对传统边缘检测算法进行了改进,采用先行自适应平滑滤波的方法,提出了一种基于3阶B样条双小波多尺度边缘检测的技术方案,利用小波的边缘检测算法对典型稀疏细胞进行边缘检测,之后综合利用图像形态学、阀值分割技术等分割出感兴趣细胞,最后采用形态学方法进行测量仿真,结果得到细胞形态参数,并对测量值与实际值进行比较,结果显示绝对误差小,满足临床要求.