Abstract:

This paper provides a review of research on deep learning algorithms in computer vision capable of automatically recognizing selected graphic patterns. The work describes the milestones in the development of computer vision from the very beginning of the idea to the currently used algorithms. The criterion for the selection of the described algorithms was based on innovation and quality improvement in relation to previously known solutions. Therefore, the work presents the most popular problems and competitions, in which participants competed for the first place in terms of the quality of the implemented models. The selected problems of computer vision that are presented in the work are: image recognition, object detection and image segmentation.