No abstract

The paper is devoted to the problem of automatic detection and recognition of license plates, the solution of which has many potential applications, from security to traffic management. The purpose of this work was to compare the methods of finding and recognizing car number plates, based on the application of deep learning algorithms, which takes into account different regional standards of car number plates, video quality, different speeds of vehicles, the location of the camera in relation to the vehicle license plate, defects of the car number plate (pollution , deformation), as well as changes in external lighting conditions. The advantages and disadvantages of localization and segmentation of car number plates on cars using image binarization, Viola–Jones and Harr methods are given. It was determined that adaptive approaches are better due to the possibility of compensating the impact of obstacles on different areas of the image, for example, the distribution of shadows due to the heterogeneity of illumination. It was determined that many methods in real algorithms rely directly or indirectly on the presence of number limits. Even if the limits are not used when the number is determined, they have the possibility to be used for further analysis. The methods of templates, image histograms, and contour analysis were compared to identify familiar features in the image (segmentation). It is shown that an effective approach for recognition of car license plates can be based on the application of the methods of Viola-Jones, Harr, the analysis of brightness histograms and the SVM method. Formulated conclusions on the effectiveness of the implementation of each of the procedures were confirmed as a result of conducting experiments with the developed software in the python 3 language using the cv2 computer vision library. The described approach makes it possible to obtain a fairly high recognition accuracy at different angles of rotation of the license plate relative to the camera. Keywords: automatic recognition, license plates, localization, normalization, segmentation, character recognition.

The article examines the process of forecasting customer outflows, which is especially important for companies that use a business model based on subscription. It was found that the outflow rate is extremely important for companies with a subscription and transactional business model, which implies regular payments to the company (banks, telecom operators, SaaS-services, etc.). For this purpose, the types, the main reasons for the outflow of customers and the parameters defined to build a predictive model using machine learning algorithms were considered. The result was the hypothesis of the reasons for the outflow of customers from sites that provide training services based on courses that are presented on-line in the Internet space. To build a model of outflow forecasting, the behavioral characteristics of students, their motivation and the structure of the courses themselves were studied. Based on the collected large array of data, their change was analyzed by a large number of parameters and the relationships between the behavioral characteristics of students, course structures and their passage were identified. A variant of the forecasting model was built, for which the accuracy of its operation was increased and the results were integrated into the customer outflow prediction module. The final list of features included more than 100 parameters, which were divided into 6 blocks. As a result, a predictive model was created using the Weibull distribution, as client behavior can be considered as a kind of survival model. To estimate the probability of customer outflow, based on the considered hypotheses, a recurrent neural network with an LSTM layer was developed, where a negative logarithmic likelihood function was used as a loss function for the Weibull distribution. As a conclusion, it was proposed to introduce a stable proactive educational business, when decisions are made not only on the basis of feelings, but also on the basis of data, comes a clearer and more sound understanding of how to improve the educational product.