Сибирский федеральный университет

Тип публикации: доклад, тезисы доклада, статья из сборника материалов конференций

Конференция: 45th International Conference on Telecommunications and Signal Processing, TSP 2022

Год издания: 2022

Идентификатор DOI: 10.1109/TSP55681.2022.9851369

Ключевые слова: convolutional neural network, image processing, letters of the russian alphabet, recognition

Аннотация: With the rapid development of computer technology, it is becoming increasingly important to have the ability to convert handwritten text into a digital version of printed text for copying, editing, or extracting data from it. The first stage of this process is the recognition of the letters of the proposed text. The article presents a new convolutional neural network (CNN) for image recognition of handwritten letters of the Russian alphabet. The resulting CNN algorithm transforms the image and recognizes the letter that is depicted on it. The final classification accuracy of the NN model on an independent dataset reaches up to 99%. © 2022 IEEE.

Журнал: 2022 45th International Conference on Telecommunications and Signal Processing, TSP 2022

Номера страниц: 272-275

Издатель: Institute of Electrical and Electronics Engineers Inc.

• Safonova A. (St. Petersburg Electrotechnical University Leti, Depart. of Automation and Control Processes, St. Petersburg, Russian Federation)
• Levkov A. (Institute of Space and Information Technologies, Siberian Federal University, Krasnoyarsk, Russian Federation)
• Kaplun D. (St. Petersburg Electrotechnical University Leti, Depart. of Automation and Control Processes, St. Petersburg, Russian Federation)

• Scopus