(25) * << * >> * Russian * English * Content * All Issues

Algorithms for training neural networks in image recognition by a uniform test

V.A. Shustov1,2
1Image Processing Systems Institute of RAS
2Samara State Aerospace University

 PDF, 116 kB

Pages: 183-189.

The article analyzes the possibility of increasing the training efficiency of a neural network that recognizes digit images. The network is configured so that all training samples are recognized. A uniform criterion of training quality is used. The considered algorithms allow both to speed up the learning process, and to reduce the number of adjustments of the neural network parameters. The last feature is important when parallelizing the learning process in cluster computing systems.

neural networks, image recognition, digit image, cluster computing.

Shustov VA. Algorithms for training neural networks in image recognition by a uniform test. Computer Optics 2003; 25: 183-189.


  1. Yezhov AA, Shumsky SA. Neurocomputing and its application in economics and business. Moscow: MEPHI Publisher; 1998. 
  2. Gorban AN, Rossiev DA. Neural networks for personal computer. Novosibirsk: “Nauka” Publisher; 1996; 276. 
  3. Kruglov VV, Borisov VV. Artificial neural networks: theory and practice. Moscow: “Goryachaya Liniya–Telekom” Publisher; 2001. 
  4. Shustov VA. Accelerating the neural network training with the selection of samples. Computer Optics 2002; 24: 160-163.

© 2009, IPSI RAS
151, Molodogvardeiskaya str., Samara, 443001, Russia; E-mail: ko@smr.ru ; Tel: +7 (846) 242-41-24 (Executive secretary), +7 (846) 332-56-22 (Issuing editor), Fax: +7 (846) 332-56-20