基于凹点检测的卵母细胞极体识别方法

为解决卵母细胞极体不明显时无法有效识别的问题,提出一种基于凹点检测的极体识别方法。结合Otsu算法和形态学操作提取细胞轮廓;通过角点检测和圆形掩膜方法搜索轮廓上的深凹点,设计自适应确定掩膜半径方法和凹凸特征参数判别深浅凹点,筛选出卵母细胞与极体粘连形成的深凹点,确定识别结果。实验结果表明,该方法在极体不明显时识别准确率达92%,耗时仅需97ms,具有鲁棒性好、自适应性强、效率高等优点,为自动化显微注射中极体识别提供了有效的解决方案。

Oocytes polar body recognition method based on concave point detection

To solve the problem that the polar body of oocytes cannot be recognized effectively when the polar body is not distinct,a polar body recognition algorithm based on concave point detection was proposed.Otsu algorithm and mathematical morphological methods were used to extract cell contour.The deep concave points on the contour were detected based on corner detection algorithm and circular mask method.The deep and shallow concave points were determined using adaptive mask radius method and concave and convex feature parameters method,and the deep concave points formed by the overlapping of oocyte and polar body were determined,and the identification results were determined.Experimental results show that the proposed algorithm costs only 97 ms and the recognition accuracy is 92%when the polar body is not obvious.It has the advantages of robustness,strong adaptability and efficientcy,which provides an effective solution for polar body recognition in automatic microinjection.