The general purpose of our research is to design and to implement image processing tools for the advancement of research in pathology and physiology. This project attempts to help the medical specialist in the diagnosis of a wide spectrum of diseases using High Performance Computing (HPC).

Image segmentation is the partitioning of an image into meaningful regions, most frequently to distinguish objects or regions of interest (“foreground”) from everything else (“background”)[1]. The current work is focused on processing digital images of human breast tissues, which present an invasive ductal carcinoma and have been treated with the immunohistochemical process.

Immunohistochemistry is a technique that allows cellular protein products to be detected in tissue components in situ by the use of the antibody: antigen reaction [2]. We have been working with two kinds of images:

1. Breast cancer images stained with ER/PR (estrogen receptor/progesterone receptor): We are working on algorithms that allow to identify and to count the nucleus of the cancer cells that react to the ER/PR.
2. Breast cancer images stained with HercepTest: We are working on algorithms that allow to segment the cell membranes, in order to determine the intensity of the staining and completeness of membrane staining.
   

   Figure 1: ER/PR

   Figure 2: HercepTest

This project is in charge of one of our researchers: Raquel Pezoa and is conducted in conjunction with a research group of Universidad de La Frontera, and is directed by Dr. Julio Lopez.

References

[1] Geoff Dougherty. Digital Image Processing for Medical Applications. Cambridge University Press. 2009.

[2]P. Phukpattaranont, P. Boonyaphiphat. Segmentation of Cancer Cells in Microscopic images using Neural Network and Mathematical Morphology, SICE-ICASE International Joint Conference 2006.