摘要
要判断一幅测得的细胞图像中是否存在癌细胞,如果仅凭经验去判断,不仅工作量大,而且准确率相对较低。文中介绍了一种基于形态学的对一幅细胞图像进行分割和识别的算法。即先对图像进行膨胀或腐蚀预处理,然后通过设置圆度阈值,计算出每一个细胞的圆度来与阈值进行比较,并提取出可疑的癌细胞。实验表明,该算法不仅大幅降低了医务人员的工作量,而且显著提高了癌细胞识别的准确率。最终检测结果的正确率达到了95%以上。
Judging only by experience whether a measured picture contains cancer cells is a huge task is relatively low in accuracy. This paper introduces a cell image segmentation and recognition algorithm based on the morphology. The image is pretreated by expansion or corrosion, and then the roundness threshold is set to be compared with the calculated roundness of each cell to find suspicious cancer cells. Experiments show that the algorithm greatly reduces the workload of medical staff, and significantly improves the accuracy of cancer cells recognition to more than 95%.
出处
《电子科技》
2016年第10期36-38,42,共4页
Electronic Science and Technology
关键词
形态学
分割
识别
圆度
细胞
morphology
image segmentation
recognition
roundness
cell
