Experimental analysis of neural network based feature extractors for cursive handwriting recognition

Artificial neural networks have been widely used in many real world applications including classification of cursive handwritten segmented characters. However, the feature extraction ability of MLP based neural networks has not been investigated properly. In this paper, a new MLP based approach such as an auto-associator for feature extraction from segmented handwritten characters is proposed. The performance of auto-associator (AA), multilayer perceptron (MLP) and multi-MLP as a feature extractor have been investigated and compared. The results and detailed analysis of our investigation are presented in the paper.