Abstract:
This article is devoted to the research and development of methods for
classifying pathologies on digital chest radiographs using two different
machine learning approaches: the eXtreme gradient boosting (XGBoost)
algorithm and the deep convolutional neural network residual network
(ResNet50). The goal of the study is to develop effective and accurate
methods for automatically classifying various pathologies detected on chest
X-rays. The study collected an extensive dataset of digital chest radiographs,
including a variety of clinical cases and different classes of pathology.
Developed and trained machine learning models based on the XGBoost
algorithm and the ResNet50 convolutional neural network using preprocessed images. The performance and accuracy of both models were
assessed on test data using quality metrics and a comparative analysis of the
results was carried out. The expected results of the article are high accuracy
and reliability of methods for classifying pathologies on chest radiographs,
as well as an understanding of their effectiveness in the context of clinical
practice. These results may have significant implications for improving the
diagnosis and care of patients with chest diseases, as well as promoting the
development of automated decision support systems in radiology.
