Experiments with various recurrent neural network architectures for handwritten character recognition

This paper reports evaluations of several neural architectures when the handwritten character recognition is approached as a problem of spectro-temporal pattern recognition. In general, neural networks specialize in learning either the spectral or temporal characteristics of patterns. However, choice of appropriate features and architectures could lead to obtaining both spectral and temporal characteristics from the handwritten character patterns. One such feature and three appropriate architectures are the focus of this paper. The results obtained during a limited set of experiments indicate a great potential for the spectro-temporal approach to be a useful contender for being a part of schemes of handwritten character recognition systems. In addition, a simple voting method is presented for collaborative character recognition using three different recognition criteria.<<ETX>>

[1]  Lokendra Shastri,et al.  Handprinted digit recognition using spatiotemporal connectionist models , 1992, Proceedings 1992 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[2]  Cris Koutsougeras,et al.  On features used for handwritten character recognition in a neural network environment , 1993, Proceedings of 1993 IEEE Conference on Tools with Al (TAI-93).

[3]  Laurene V. Fausett,et al.  Fundamentals Of Neural Networks , 1994 .

[4]  Cris Koutsougeras,et al.  Experiments with Kohonen's learning vector quantization in handwritten character recognition systems , 1994, Proceedings of 1994 37th Midwest Symposium on Circuits and Systems.

[5]  Cris Koutsougeras,et al.  Experiments with various neural architectures for handwritten character recognition , 1994, Proceedings of 1994 37th Midwest Symposium on Circuits and Systems.