Manual processing of medical images obtained using various imaging methods can take a long time for a specialist, therefore, the task of automatic image segmentation is now urgent in medical diagnostics in order to speed up the time between conducting research and making a diagnosis. With the development of computing power and the availability of various machine learning libraries, convolutional neural networks that have proven themselves in image segmentation tasks are increasingly being used to solve it. In this paper, we consider the task of segmenting human lung images into lobes using a convolutional neural network. Images obtained by computed tomography. The paper presents the network architecture, prepared a data set, describes the metrics used, conducted experiments, and obtained results.