Summary
In this article, we propose an automated construction of knowledge based artificial neural networks (KBANN) for the recognition of restricted sets of handwritten words or characters. The features that better describe the chosen vocabulary are first selected, according to the characteristics of the used script, language and lexicon. Then, ideal samples of lexicon elements (words or characters) are submitted to a feature extraction module to derive their description using the chosen primitives. The analysis of these descriptions generates a symbolic knowledge base reflecting a hierarchical classification of the words (or characters). The rules are then translated into a multilayer neural network by determining precisely its architecture and initializing its connections with specific values. This construction approach reduces the training stage, which enables the network to reach its final topology and to generalize. The proposed method has been tested on the automated construction of neuro-symbolic classifiers for two Arabic word lexicons.
See the full content of this document
Extract
Toward a Generic Hybrid Neural System for Handwriting Recognition: An Application to Arabic Words
1. Introduction
In recent years, there has been an explosive growth in the successful use of hybrid intelligent systems in many diverse areas [10, 21] such as robotics, medical diagnosis, speech/ natural language understanding, fault diagnosis of industrial equipment, monitoring/ control of manufacturing processes and financial applications.Given the number and variety of methods offered for pattern recognition, for instance, there is no single method that can be called the best. Each approach has strengths and weaknesses, good and bad ideas. One way to take advantage of this variety is to build multiple sources of information based systems. This direction is given more attention in pattern recognition and more work is being done, especially for handwriting recognition applications. The reported results show the efficiency of such techniques including hybrid approaches and multiple classifier schemes [1, 4, 16, 17, 20].The multiple classifier approach is defined as a system consisting of a set of classifiers a...See the full content of this document
Sponsored links
