Research Databases Bronchial Images

Microscopical Image Database (Bronchial cytology)

This database contains 8 color microscopic images from bronchial cytology. A ground truth is also available for each image in three classes (nucleus, cytoplasm, and background pixels). Pixels have a label specifying their classes (2: nucleus, 1: cytoplasm, 0: background). Figures below illustrate the purpose of the image processing process: the classification of pixels (left: original image, center: ground truth, right: segmented image).
Image Ground truth Segmentation result

You can access to this database by the following link. Download the Bronchial Image Database.

You are free to use the dataset for non-commercial research and educational purposes. In exchange, we request only that you cite our IJRA paper:

@article{rin-Meurie-2005-1,
author={C. Meurie and O. Lezoray and C. Charrier and A. Elmoataz},
title={Combination of multiple pixel classifiers for microscopic image segmentation},
journal={IJRA (IASTED International Journal of Robotics and Automation)},
volume={20},
year={2005},
pages={63-69},
number={2},
note={Special issue on Colour Image Processing and Analysis for Machine Vision,
ISSN 0826-8185}
}