Theses to the question of the software methodology for the application of the technology of the recognition neural network device

The article deals with the general problem of the problem of localization and classification of objects in raster images. The task of localization as a technology of iterative recognition of road signs as a typical object for demonstrating recognition methodology. At the moment, there are no universal methods and technologies that can ideally solve the problem of localizing and classifying objects in an image. The most popular and relevant methods for localization and classification of objects in the image were analyzed. An algorithm for recognition technology has been developed and presented. The results of the program work are demonstrated. The target map of using the algorithmic method is generalized. To solve the problems posed in this study, it was decided to use HOG & SVM descriptors in the localization module, and transfer the localized image to the CNN that implements the classification module. An algorithm for developing an own model for designing recognition technology based on the methodology for using a neural network apparatus in software is considered and justified.