Medical digital images based on the basis model

History

2024: Neuromodeling Lab Launch

On May 27, 2024, Donskoy State Technical University (DSTU) announced the formation of the Medical Digital Images Based on the Basis Model laboratory. We are talking about a neuromodeling site for conducting research on biological samples using artificial intelligence.

Laboratory specialists, in particular, will develop an algorithm for the basic model of a digital platform for storing, processing and analyzing medical images. In addition, a neural network will be created to process changes in brain cells: it will help identify possible pathologies. According to Yevgeny Kirichenko, head of the Department of Bioengineering at DSTU, the research area is a network of gap contacts between astrocytes in normal and tumor conditions.

source = DSTU

Our main task is to take pictures of rat brain astrocytes normally and human astrocytes during tumor transformation. The neural network will determine whether the astrocyte is normal or pathological - for example, due to tumor growth or Alzheimer's disease, says Kirichenko.

As a training set of data for the neural network, medical images marked by specialists will serve: these images will be uploaded to ChatGPT. In addition, within the framework of the laboratory, it is planned to create a digital pathologist with ChatGPT in collaboration with the Regional Children's Clinical Hospital. The project is aimed at creating a digital model of clinically significant biological fluids of humans and animals in normal and pathological conditions. The work will use impersonal cytology images indicating the gender, age and clinical situation of the patient. The use of AI will allow you to identify patterns in data and make decisions based on them.

The laboratory is created jointly with the St. Petersburg Polytechnic University of Peter the Great. It will enter the campus of Blue Sky Research digital laboratories. The project partner is the neural network architecture company Vizhntekh LLC^[1]

Notes