Printed Arabic words recognition using genetic algorithm

Abstract EN

The automatic recognition of text on scanned images has enabled many application such as searching for words in large volumes of documents, automatic sorting of postal mail, and convenient editing of previously printed documents.

In this paper, a Printed Arabic word recognition system using Wavelet Transform, and Genetic Algorithms (GA) is proposed.

The system consist three steps.

In first step pre-processing are applied on the input image.

Secondly features are extracted, which will be taken as the input to the genetic algorithm.

In feature extraction stage the feature vector extracted by using 1-level linear wavelet decomposition technique and only the approximation are taken as the input to the (GA).

While the third step is a classification which is carried out by GA.

The proposed approaches are tested on a ten Arabic word.

The recognition rate is 90%.