The thesis aims to explore the possibility to use machine learning and neural networks to reveal a specific gene mutation from the visual appearance of the cancer tissue. 
 
The student will design, train, and evaluate a neural network for detecting mutations from the whole-slide histopathological images of colorectal cancer tissue. The student will implement the necessary preprocessing of images and other sources, provided by the Masaryk Memorial Cancer Institute, for usage in the machine learning pipeline. Evaluation of the model will include visualizations shown as overlays on top of the background images. 

The student will implement the proposed methods and associated tools, and integrate them with the tools developed within the RationAI group. The neural network model will be developed, trained, and evaluated within the cloud infrastructure of the RationAI group.

The thesis text will contain a brief overview of approaches in this area, a description of the methodology, and an explanation of performed experiments and obtained results.