Biomedical image analysis

Digital pathology

The objective of this research is to develop automated or semi-automated methods for computer-assisted classification of histology images, in particular for the following tasks:

Discrimination between normal and abnormal tissue;
Classification of tissue sub-types such as epithelium, stroma, lymphocytes and necrosis.

We are also studying the effects of colour normalisation on automated classification procedures.

In the picture: different tissue subtypes from colorectal cancer. From top to bottom: epithelium, stroma, lymphocytes and debris. From Kather J.N. et al. Collection of textures in colorectal cancer histology. See also:

Kather, J.N. et al. Multi-class texture analysis in colorectal cancer histology (2016) Scientific Reports, 6, art. no. 27988
Bianconi, F., Kather, J.N. and Reyes-Aldasoro, C.C. Evaluation of Colour Pre-processing on Patch-Based Classification of H&E-Stained Images (2019) Lecture Notes in Computer Science, 11435, pp. 56-64.