Skip to main content
high contrast version
pl polski menu
"

Web Content Display Web Content Display

Skip banner

Grant Recruitment

Seminars

ID.UJ Competitions

For PhD Students

Web Content Display Web Content Display

Implementation Doctorate Programme II - artificial intelligence of the Doctoral School of Exact and Natural Sciences of the Jagiellonian University for 2020, part 1


Information on endowment from the state budget or a state special purpose fund
The project is financed by the state budget.

Name of the programme or fund
IMPLEMENTATION DOCTORATE PROGRAMME II - ARTIFICIAL INTELLIGENCE

Name of the project
Implementation Doctorate Programme II - artificial intelligence of the Doctoral School of Exact and Natural Sciences of the Jagiellonian University for 2020

Value of the endowment
PLN 309 522.34

Total investment cost
PLN 309 522.34

Brief description of the project
The goal of the project is to develop techniques for analyzing biomedical images obtained with the help of High Content Screening (HCS) technology. There is a number of challenges involving analyzing biomedical imagery, i.e. relatively little training data and subtle differences between data classes (between phenotypic changes in cells caused by the processes in chemical compounds). Moreover, in the case of biomedical data, a very important aspect is the interpretability and explainability of the results of automatic analysis, e.g. due to EU law obliging decision support systems to be self-explaining. Therefore, it is very important that the developed algorithms can be used on data coming from various studies and centers in order to neutralize differences resulting from a different thickness in which the tissue is cut or its arrangement.

The listed challenges constitute a research problem that can be solved by:

  • The use of algorithms based on deep neural networks;
  • Development of standardization methods for the analyzed medical data;
  • The use of XAI (Explainable Artificial Intelligence) methods to explain and authenticate the developed models;
  • The use of knowledge transfer and domain adaptation methods to research material from different parts of the body.

Expected outcome:

  • Obtaining a representation of the examined medical images that is resistant to differences in data received from different studies and centers;
  • Preparation of image quality testing procedure;
  • Preparation of a procedure for identifying images that negatively affect the learning process of the algorithm (e.g. containing artifacts);
  • Development of methods for detecting changes occurring at the cellular level responsible for developing the disease;
  • Development of methods for recognizing and quantifying tissue features, i.e. cell type and size;
  • Development of a method for determining ROI ( Region of Interest) containing pathological changes often occurring in a small area of tissue.

Schedule for the implementation of scientific work