Color region annotation for microvessel density estimation Conference Paper uri icon


  • In this paper, we proposed a method for microvessel density estimation based on color region annotation. First, the pericytes and endothelial cells of the microvessel are modeled in the RGB color space by minimum error thresholding. Second, the connected component labeling method is used to extract microvessels in the neighboring red and brown color regions. Third, the microvessel density is estimated by the number of detected endothelial pixels in the image. Finally, the discrimination ability of the number of endothelial cells and the color contrast of microvessel is validated on the classification of both positive and negative images. The experimental results demonstrate that the proposed method is quite promising. © 2012 IEEE.

author list (cited authors)

  • Mete, M., Chen, J., Wen, Q., & Liu, X.

publication date

  • January 1, 2013 11:11 AM