АNALYSIS OF APPROACHES TO SOLVING THE PROBLEM OF PICTURE RECOGNITION USING ARTIFICIAL INTELLIGENCE
DOI:
https://doi.org/10.26906/SUNZ.2023.1.096Keywords:
neural network, perceptron, weight paradigm, cnn, rnn, adaptive resonance theoryAbstract
Topicality. Text recognition on images (optical character recognition) is one of the areas of image recognition, the task of which is to translate images of handwritten, typewritten or printed text into text data that is used to represent characters on a computer (for example, in a text editor). . Recognizing text on images is an important task of machine learning, as it allows you to organize convenient interaction with data: editing, analysis, searching for words or phrases, etc. Nevertheless, the creation of an application in this field remains a creative task and requires additional research in connection with the specific requirements for resolution, speed, recognition reliability and memory capacity, which are characterized by each specific task. In the work, the algorithm for creating and learning the recognition of handwritten symbols of a neural network was considered, and the types of learning and classification of neural networks were also analyzed. The goal of this work there are recommendations and approaches to the selection of types of neural networks, methods and their training. The object of research is the task of image recognition of handwritten symbols. The object of research image recognition task. The subject of research recognition of images of text or symbols based on artificial intelligence. Results. the paper analyzed the types, training and classification of neural networks for creating recognition of handwritten symbols using a neural network. Conclusions. Currently, there is a fairly large number of innovative companies on the market that are engaged in image recognition using neural network learning system technologies. It is known for certain that they achieved image recognition accuracy of around 95% when using a database of 10,000 images. However, all achievements refer to static images, the situation is not clear-cut with dynamic images. Therefore, further research in the field of image recognition using a neural network is still being investigated, it is relevant now.Downloads
References
Нейронні мережі: їх застосування, робота [Electronic resource] – URL: https://www.poznavayka.org/
Neural Networks Modeling and Control: Applications for Unknown Nonlinear Delayed Systems in Discrete Time Authors: Jorge Rios, Alma Alanis, Nancy Arana-Daniel, Carlos Lopez-Franco -2020. – 158 p.
Згорткова нейронна мережа – просте пояснення CNN та її застосування [Electronic resource] – URL: https://evergreens.com.ua/ua/articles/cnn.html
Рекурентні нейронні мережі [Electronic resource] – URL: https://studfile.net/preview/7653871/page:9/
Навчання з підкріпленням у машинному навчанні [Electronic resource] – URL: https://goo.su/i0gLlz9
Y. LeCun, L. Bottou, G. Orr and K. Muller: Efficient BackProp, in Orr, G. and Muller K. (Eds) // Neural Networks: Tricks of the trade. – 2008. – 200 p.
Y. LeCun and Y. Bengio: Convolutional Networks for Images, Speech, and Time-Series, in Arbib, M. A. (Eds) // The Handbook of Brain Theory and Neural Networks. – 2005. – 150 p.
Neural Networks and Deep Learning: A Textbook by Charu C. Aggarwal – 2018. - 520 p.
