Geodesic
Achieving statistical dependence of the CNN response on the input data distortion for OCR problem
Informacionnye tehnologii i vyčislitelnye sistemy, no. 4 (2019), pp. 94-101.

Voir la notice de l'article provenant de la source Math-Net.Ru

The paper proposes an approach to training a convolutional neural network using information on the level of distortion of input data. The learning process is modified with an additional layer, which is subsequently deleted, so the architecture of the original network does not change. OCR of data based on the MNIST dataset distorted with Gaussian blur using LeNet5 architecture network is considered. This approach does not have quality loss of the network and has a significant error-free zone in responses on the test data which is absent in the traditional approach to training. The responses are statistically dependent on the level of input image’s distortions and there is a presence of a strong relationship between them.
Keywords: Convolutional neural networks, pattern recognition, machine learning, distortion, Gaussian blur
Mots-clés : OCR, MNIST. 
@article{ITVS_2019_4_a8,
     author = {I. M. Janiszewski and V. V. Arlazarov and D. G. Slugin},
     title = {Achieving statistical dependence of the {CNN} response on the input data distortion for {OCR} problem},
     journal = {Informacionnye tehnologii i vy\v{c}islitelnye sistemy},
     pages = {94--101},
     publisher = {mathdoc},
     number = {4},
     year = {2019},
     language = {en},
     url = {http://geodesic.mathdoc.fr/item/ITVS_2019_4_a8/}
}
TY  - JOUR
AU  - I. M. Janiszewski
AU  - V. V. Arlazarov
AU  - D. G. Slugin
TI  - Achieving statistical dependence of the CNN response on the input data distortion for OCR problem
JO  - Informacionnye tehnologii i vyčislitelnye sistemy
PY  - 2019
SP  - 94
EP  - 101
IS  - 4
PB  - mathdoc
UR  - http://geodesic.mathdoc.fr/item/ITVS_2019_4_a8/
LA  - en
ID  - ITVS_2019_4_a8
ER  - 
%0 Journal Article
%A I. M. Janiszewski
%A V. V. Arlazarov
%A D. G. Slugin
%T Achieving statistical dependence of the CNN response on the input data distortion for OCR problem
%J Informacionnye tehnologii i vyčislitelnye sistemy
%D 2019
%P 94-101
%N 4
%I mathdoc
%U http://geodesic.mathdoc.fr/item/ITVS_2019_4_a8/
%G en
%F ITVS_2019_4_a8
I. M. Janiszewski; V. V. Arlazarov; D. G. Slugin. Achieving statistical dependence of the CNN response on the input data distortion for OCR problem. Informacionnye tehnologii i vyčislitelnye sistemy, no. 4 (2019), pp. 94-101. http://geodesic.mathdoc.fr/item/ITVS_2019_4_a8/