Main > Acta Optica Sinica >  Volume 34 >  Issue 1 >  Page 115001 > Article
  • Abstract
  • Abstract
  • View Summary
  • Figures (0)
  • Tables (0)
  • Equations (0)
  • References (18)
  • Get PDF(in Chinese)
  • Paper Information
  • Received: Jul. 9, 2013

    Accepted: --

    Posted: Jan. 1, 2014

    Published Online: Jan. 2, 2014

    The Author Email: Tao Guan (gt_mike2003@126.com)

    DOI: 10.3788/aos201434.0115001

  • Get Citation
  • Copy Citation Text

    Guan Tao, Zhou Dongxiang, Liu Yunhui. Color Optical Microscopic Cell Image Segmentation Based on Color Difference Vector Field[J]. Acta Optica Sinica, 2014, 34(1): 115001

    Download Citation

  • Category
  • Machine Vision
  • Share
Acta Optica Sinica, Vol. 34, Issue 1, 115001 (2014)

Color Optical Microscopic Cell Image Segmentation Based on Color Difference Vector Field

Tao Guan1,*, Dongxiang Zhou1, and Yunhui Liu2

Author Affiliations

  • 1[in Chinese]
  • 2[in Chinese]

Abstract

Cell image segmentation is one of the hot topics in medical image processing. Most of the classical algorithms for cell image segmentation are based on grayscale images, which results in loss of color information in images. Based on analyzing the characteristics of the color cell images, we present a color difference vector field to model the color feature of cell images. In the color difference vector field, the difference between cell region and non-cell region is more distinct compared with other classical color spaces, such as HSV, YIQ and CIEL*a*b* spaces. Furthermore, this method is more robust for a large number of cell images. Based on the color difference vector field, a sequential match method is proposed for segmentation of cell images. In order to obtain more accurate results, the color difference strength is used to refine the segmentation results. Various color cell images containing overlapped cells have been tested to show the validity and effectiveness of the proposed method. The accuracy of the proposed method reaches 95.2%, which is higher than that of the RGVF Snake method.

keywords

Please Enter Your Email: