Publication:
Combining additive input noise annealing and pattern transformations for improved handwritten character recognition

Loading...
Thumbnail Image
Identifiers
Publication date
2014-12-15
Defense date
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Elsevier
Impact
Google Scholar
Export
Research Projects
Organizational Units
Journal Issue
Abstract
Two problems that burden the learning process of Artificial Neural Networks with Back Propagation are the need of building a full and representative learning data set, and the avoidance of stalling in local minima. Both problems seem to be closely related when working with the handwritten digits contained in the MNIST dataset. Using a modest sized ANN, the proposed combination of input data transformations enables the achievement of a test error as low as 0.43%, which is up to standard compared to other more complex neural architectures like Convolutional or Deep Neural Networks. © 2014 Elsevier Ltd. All rights reserved.
Description
Keywords
artificial neural-networks, back propagation, handwritten text recognition, MNIST
Bibliographic citation
Alonso-Weber, J.M., Sesmero, M.P., Sanchis, A. (2014). Combining additive input noise annealing and pattern transformations for improved handwritten character recognition. Expert Systems with Applications, 41(18), pp. 8180-8188