This paper validates a recognition system via multiwavelet neural network as feature extractor and classifier. It investigates the relevance of each sub-band image in the recognition process. An experiment to verify the efficiency of the multiwavelet was performed omitting the feature extraction step. Results show that information about the relevant image features are evenly distributed in all sub-band images of multiwavelet coefficients and that multiwavelet neural network are promising feature extractors and classifiers. Numerals from the NIST database were used for evaluation of the system tested.
Performance Analysis of Handwritten Numerals Recognition Based on Multiwavelet Neural Network
2008 Congress on Image and Signal Processing ; 4 ; 380-384
2008-05-01
440434 byte
Conference paper
Electronic Resource
English
British Library Conference Proceedings | 1995
|Evaluation of Codes and Primitives: Recognition of Unconstrained Handwritten Numerals
British Library Conference Proceedings | 1995
|