计算机工程与应用 ›› 2019, Vol. 55 ›› Issue (11): 192-198.DOI: 10.3778/j.issn.1002-8331.1803-0268

• 图形图像处理 • 上一篇    下一篇

DNA机器人在AFM图像中的分割和识别

童  麟1,2,韩越兴1,3,小长谷明彦3,4   

  1. 1.上海大学 计算机工程与科学学院,上海 200444
    2.中国科学院大学 上海高等研究院,上海 200000
    3.东京工业大学 综合理工研究科,东京 226-8502
    4.国立情报学研究所,东京 101-8430
  • 出版日期:2019-06-01 发布日期:2019-05-30

Segmentation and Recognition of DNA Robots in AFM Images

TONG Lin1,2, HAN Yuexing1,3, Akihiko Konagaya3,4   

  1. 1.School of Computer Engineering and Science, Shanghai University, Shanghai 200444, China
    2.Shanghai Advanced Research Institute, University of Chinese Academy of Sciences, Shanghai 200000, China
    3.Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology, Tokyo 226-8502, Japan
    4.National Institute of Informatics, Tokyo 101-8430, Japan
  • Online:2019-06-01 Published:2019-05-30

摘要: DNA机器人是一种由DNA大分子通过折纸技术制作的纳米级别的机器人,可以用于癌症的诊断和治疗。为了对DNA机器人形态进行研究,研究人员利用原子力显微镜(Atomic Force Microscope,AFM)拍摄出机器人的AFM图像。针对AFM图像多噪声及被观察的DNA机器人大规模重叠的特点,从除噪、分割、识别三个过程展开,提出了一种基于多尺度改进的分水岭AFM图像切割算法,以傅里叶描述子、曲率尺度空间描述符和Hu矩作为特征,实现DNA纳米机器人的分类和识别。

关键词: 原子力显微镜(AFM)图像, 分水岭算法, DNA机器人, 物体识别, 噪声消除

Abstract: DNA robot is a type of nano robots made from DNA macromolecules with the origami technology. DNA robots can be used for the diagnosis and treatment of the cancer. To research the DNA robots, Atomic Force Microscope(AFM) is used to take the photographs of the DNA robots. Since there are many noises in the AFM images and overlaps among the DNA robots, removing noises, segmentation, and recognition are processed. The watershed based on multi-scale technique is improved for the segmentation of DNA robots. DNA robots are classified and recognized with the Fourier descriptors, Curvature Scale Space(CSS) method and Hu moment.

Key words: Atomic Force Microscope(AFM) image, watershed algorithm, DNA robots, object recognition, noise elimination