This article examines the effectiveness of gender and age recognition using a neural network. The search and analysis of existing methods of image recognition and selection of characteristics from them are performed, and similar studies are described. The description of the used model, the principle of its operation, training parameters, and the reason for choosing this model is produced. An experiment is being conducted on the effectiveness of gender and age recognition based on chest x-rays. To conduct the experiment, a neural network model is created based on the existing one. The obtained results are analyzed. Conclusions are drawn on the experiment and the possibility of further research.