CAJ | 학술논문

针对斑马鱼视网膜细胞图像分割时细胞粘连比较严重的情况，采用了边界距离变换和模拟浸水原理的分水岭变换相结合的方法。首先采用距离变换的方法提取细胞区域的局部极小值作为种子点，合并距离过小的错误种子点，然后使用分水岭算法进行分割。实验结果表明，该方法能够有效地分割斑马鱼视网膜细胞图像，成功地解决了分水岭变换中粘连细胞的过分割问题，且计算速度快。
침대반마어시망막세포도상분할시세포점련비교엄중적정황，채용료변계거리변환화모의침수원리적분수령변환상결합적방법。수선채용거리변환적방법제취세포구역적국부겁소치작위충자점，합병거리과소적착오충자점，연후사용분수령산법진행분할。실험결과표명，해방법능구유효지분할반마어시망막세포도상，성공지해결료분수령변환중점련세포적과분할문제，차계산속도쾌。
For image segmentation cases of zebra fish retinal cells in serious overlapped, the paper uses the boundary distance transform and the principle of immersion simulation method of combining watershed transform. The basic principle is extracting the local minimum values of ceils as a regional seeds point by using distance transformation. It deletes the error seed points which are too close, and then divided by the watershed algorithm. Experimental results show that the method can effectively split the image of zebra fish retinal cells. The method has successfully solved the over-segmentation problem of watershed transformation in cells adhe- sion, and the calculation is fast.