中国邮电高校学报(英文) ›› 2018, Vol. 25 ›› Issue (5): 83-92.doi: 10.19682/j.cnki.1005-8885.2018.0027

• Others • 上一篇    下一篇

Segmentation of overlapping cervical nuclei based on the identification

赵晶1,谢怡宁2,卢玉1,何勇军2   

  1. 1. 哈尔滨理工大学计算机科学与技术学院
    2. 哈尔滨理工大学
  • 收稿日期:2018-01-10 修回日期:2018-07-29 出版日期:2018-10-18 发布日期:2018-10-18
  • 通讯作者: 何勇军 E-mail:Holywit@163.com
  • 基金资助:
    中国国家自然科学基金;中国国家自然科学基金;中国国家自然科学基金;高等学校博士学科点专项科研基金;哈尔滨理工大学青年拔尖创新人才培养计划

Segmentation of overlapping cervical nuclei based on the identification

  • Received:2018-01-10 Revised:2018-07-29 Online:2018-10-18 Published:2018-10-18
  • Supported by:
    National Nature Science Foundation of China;National Nature Science Foundation of China;National Nature Science Foundation of China

摘要: Image segmentation directly determines the performance of automatic screening technique. However, there are overlapping nuclei in nuclei images. It raises a challenge to nuclei segmentation. To solve the problem, a segmentation method of overlapping cervical nuclei based on the identification is proposed. This method consists of three stages: classifier training, recognition and fine segmentation. In the classifier training, feature selection and classifier selection are used to obtain a classifier with high recognition rate. In the recognition, the outputs of the rough segmentation are classified and processed according to their labels. In the fine segmentation, the severely overlapping nuclei are further segmented based on the prior knowledge provided by the recognition. Experiments show that this method can accurately segment overlapping nuclei.

关键词:

overlapping nuclei segmentation, pits detection, deep learning, segmentation strategy

Abstract: Image segmentation directly determines the performance of automatic screening technique. However, there are overlapping nuclei in nuclei images. It raises a challenge to nuclei segmentation. To solve the problem, a segmentation method of overlapping cervical nuclei based on the identification is proposed. This method consists of three stages: classifier training, recognition and fine segmentation. In the classifier training, feature selection and classifier selection are used to obtain a classifier with high recognition rate. In the recognition, the outputs of the rough segmentation are classified and processed according to their labels. In the fine segmentation, the severely overlapping nuclei are further segmented based on the prior knowledge provided by the recognition. Experiments show that this method can accurately segment overlapping nuclei.

Key words: overlapping nuclei segmentation, pits detection, deep learning, segmentation strategy