KFU. Card publication. An Effect of Binarization on Handwritten Digits Recognition by Hierarchical Neural Network

KAZAN
FEDERAL UNIVERSITY

- RUS
- 中文
- ES

AN EFFECT OF BINARIZATION ON HANDWRITTEN DIGITS RECOGNITION BY HIERARCHICAL NEURAL NETWORK

Form of presentation

Articles in international journals and collections

Year of publication

2022

��

��

Authors, employees of KFU

Kayumov Zufar Damirovich, author

Mosin Sergey Gennadevich, author

Tumakov Dmitriy Nikolaevich, author

Bibliographic description in the original language

Kayumov Z, Tumakov D, Mosin S., An Effect of Binarization on Handwritten Digits Recognition by Hierarchical Neural Network//Lecture Notes in Networks and Systems. - 2022. - Vol.300 LNNS, Is.. - P.94-106.

Annotation

The influence of binarization on the recognition of handwritten digits has been investigated. The main methods of converting images from grayscale to black and white images are analyzed. The binarization process is applied to images of handwritten digits. Both the Convolutional neural network (CNN) and the multilayer perceptron (MLP) architectures are proposed for an automated recognizing the handwritten digits. The CNN and MLP are trained on the raw data of images. Essential characteristics are obtained by the method of feature extraction from binarized images and then are used for the CNN and MLP training. The influence of the binarization threshold on network recognition errors is analyzed. A comparison of the recognition accuracy by the neural networks for different algorithms is carried out. It was estimated that the minimum error of combining the proposed networks does not exceed 99.86%. A new approach to the design of a hierarchical network composed of a network trained on basic images of digits and images of digits obtained by binarization is proposed. The hierarchical neural networks with an accuracy of 99.61% on the MNIST dataset are proposed.

Keywords

Handwritten digits, Recognition, CNN, Hierarchical neural network, Binarization, MNIST

The name of the journal

Lecture Notes in Networks and Systems

URL

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85115624156&doi=10.1007%2f978-3-030-84760-9_9&partnerID=40&md5=2fde4b83bb5711f32d33668e5fdc21c9

Please use this ID to quote from or refer to the card

https://repository.kpfu.ru/eng/?p_id=258363&p_lang=2

Full metadata record

Field DC	Value	Language
dc.contributor.author	Kayumov Zufar Damirovich	ru_RU
dc.contributor.author	Mosin Sergey Gennadevich	ru_RU
dc.contributor.author	Tumakov Dmitriy Nikolaevich	ru_RU
dc.date.accessioned	2022-01-01T00:00:00Z	ru_RU
dc.date.available	2022-01-01T00:00:00Z	ru_RU
dc.date.issued	2022	ru_RU
dc.identifier.citation	Kayumov Z, Tumakov D, Mosin S., An Effect of Binarization on Handwritten Digits Recognition by Hierarchical Neural Network//Lecture Notes in Networks and Systems. - 2022. - Vol.300 LNNS, Is.. - P.94-106.	ru_RU
dc.identifier.uri	https://repository.kpfu.ru/eng/?p_id=258363&p_lang=2	ru_RU
dc.description.abstract	Lecture Notes in Networks and Systems	ru_RU
dc.description.abstract	The influence of binarization on the recognition of handwritten digits has been investigated. The main methods of converting images from grayscale to black and white images are analyzed. The binarization process is applied to images of handwritten digits. Both the Convolutional neural network (CNN) and the multilayer perceptron (MLP) architectures are proposed for an automated recognizing the handwritten digits. The CNN and MLP are trained on the raw data of images. Essential characteristics are obtained by the method of feature extraction from binarized images and then are used for the CNN and MLP training. The influence of the binarization threshold on network recognition errors is analyzed. A comparison of the recognition accuracy by the neural networks for different algorithms is carried out. It was estimated that the minimum error of combining the proposed networks does not exceed 99.86%. A new approach to the design of a hierarchical network composed of a network trained on basic images of digits and images of digits obtained by binarization is proposed. The hierarchical neural networks with an accuracy of 99.61% on the MNIST dataset are proposed.	ru_RU
dc.language.iso	ru	ru_RU
dc.subject	Handwritten digits	ru_RU
dc.subject	Recognition	ru_RU
dc.subject	CNN	ru_RU
dc.subject	Hierarchical neural network	ru_RU
dc.subject	Binarization	ru_RU
dc.subject	MNIST	ru_RU
dc.title	An Effect of Binarization on Handwritten Digits Recognition by Hierarchical Neural Network	ru_RU
dc.type	Articles in international journals and collections	ru_RU