Detecting Graphic Objects Using an Artificial Neural Network

The article considers an approach to editing graphic objects (for example, road signs) using artificial layered convolutional neural networks. A mobile road laboratory “TRASSA” was used as engineering equipment for collecting information. The experimental results demonstrate high accuracy of identification of road signs in different environments. An object recognition approach using deep machine learning is described, in which a convolutional neural network is used and software for identifying road signs is developed. The suggested approach makes it possible to discover the existence of detectable objects in data flows and to identify them on this basis.