Project Manager: Makhanov N.M.
Funding source: GF of young scientists for the project “Zhas Galym”
Project objective: To develop an architecture for diagnosing Covid-19 and other pathologies based on lung X-rays using Federated Learning and small-example learning methods.
Years of implementation: 2024 – 2026
Funding amount: 29,858,850 tenge
Federated learning with multi-stage learning will advance scientific knowledge in the field of chest medical image classification. The project will also create a small web prototype that will be used to scale and commercialize the technology. The results of the project “Development of Federated Learning Algorithms and Integration of Small Sample Learning Methods for Classification of Chest Diseases in X-ray Images.” are primarily targeted at various private and public medical organizations involved in the detection of chest diseases using X-ray images. They will be of broader interest to priority areas of scientific development: Intellectual Potential of the Country, Life and Health Sciences, National Security and Defense, as well as for scientists and institutions involved in the design and development of AI research applied to medical imaging. The research results include publications in at least two journals in the top three quartiles of the Web of Science impact factor or having a CiteScore percentile of at least 50 in Scopus in the fields of computer vision, medicine and artificial intelligence. A workshop will be organized at AITU with the participation of scientists and industry representatives to disseminate the main research results. A patent application for the new arrangement will also be considered after the success of the planned tests. The research results and data will be archived on the servers provided by AITU. To disseminate the research results, increase the likelihood of implementation and commercialization of the project, a web page will be created on the AITU website, which will contain brief information about the project, including relevance, objective, expected and achieved results, name and surname of the research team members. |