Combining additive input noise annealing and pattern transformations for improved handwritten character recognition |
| |
Affiliation: | 1. Anesthesia and Critical Care Research Group, Department of Clinical Medicine, UiT, The Arctic University of Norway, 9037 Tromsø, Norway;2. Division of Surgical Medicine and Intensive Care, University Hospital of North Norway, 9038 Tromsø, Norway;3. Department of Research and Education, Norwegian Air Ambulance Foundation, 1441 Drøbak, Norway;1. Department of Mathematics, Kyoto Sangyo University, Motoyama, Kamigamo, Kita-Ku, Kyoto, 603-8555, Japan;2. Department of Mathematics, Kansai University, Suita, Osaka 564-8680, Japan;1. Madurai Kamaraj University, Madurai 625 021, India;2. Sri Meenakshi Govt. Arts College for Women(A), Madurai 625 002, India;1. Centre of Image and Signal Processing, Faculty of Computer Science and Information Technology, University of Malaya, 50603 Kuala Lumpur, Malaysia;2. School of Computing, National University of Singapore, Singapore;1. School of Control and Computer Engineering, North China Electric Power University, Beijing 102206, PR China;2. Technology Center of Software Engineering, Institute of Software, Chinese Academy of Sciences, Beijing 100190, PR China;1. University of Belgrade, Technical Faculty in Bor, VJ 12, Bor, Serbia;2. University of Belgrade, Faculty of Mining and Geology, Djusina 7, Belgrade, Serbia |
| |
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. |
| |
Keywords: | Artificial Neural Networks Back Propagation MNIST Handwritten text recognition |
本文献已被 ScienceDirect 等数据库收录! |
|