Browsing by Author "Chikhi, Salim"
Now showing 1 - 3 of 3
Results Per Page
Sort Options
Item A novel fuzzy approach for handwritten arabic character recognition(Springer, 2015) Kef, Maamar; Chergui, Leila; Chikhi, SalimThe aim of our work is to present a new method based on structural characteristics and a fuzzy classifier for off-line recognition of handwritten Arabic characters in all their forms (beginning, end, middle and isolated). The proposed method can be integrated in any handwritten Arabic words recognition system based on an explicit segmentation process. First, three preprocessing operations are applied on character images: thinning, contour tracing and connected components detection. These operations extract structural characteristics used to divide the set of characters into five subsets. Next, features are extracted using invariant pseudo-Zernike moments. Classification was done using the Fuzzy ARTMAP neural network, which is very fast in training and supports incremental learning. Five Fuzzy ARTMAP neural networks were employed; each one is designed to recognize one subset of characters. The recognition process is achieved in two steps: in the first one, a clustering method affects characters to one of the five character subsets. In the second one, the pseudo-Zernike features are used by the appropriate Fuzzy ARTMAP classifier to identify the character. Training process and tests were performed on a set of character images manually extracted from the IFN/ENIT database. A height recognition rate was reported.Item Combining neural networks for arabic handwriting recognition(IEEE, 2011) Chergui, Leila; Kef, Maâmar; Chikhi, SalimCombining classifiers is an approach that has been shown to be useful on numerous occasions when striving for further improvement over the performance of individual classifiers. In this paper we present an off-line Multiple Classifier System (MCS) for Arabic handwriting recognition. The MCS combine two individual recognition systems based on Fuzzy ART network used for the first time in Arabic OCR, and Radial Basis Functions. We use various feature sets based on Hu and Zernike Invariant moments. For deriving the final decision, different combining schemes are applied. The best combination ensemble has a recognition rate of 90,1 %, which is significantly higher than the 84,31% achieved by the best individual classifier. To demonstrate the high performance of the classification system, the results are compared with three research using IFN/ENIT database.Item New hybrid arabic handwriting recognizer(IEEE, 2012) Chergui, Leila; Kef, Maâmar; Chikhi, SalimRecently, there is a popular belief that classifier combination of different architecture could complement each other for improving results performance. In this paper we introduce a framework to combine results of multiple classifiers for offline Arabic handwriting recognition, by introducing a new scheme of combination of Multi Layer Perceptron and ART1 networks. Besides using two different recognition architectures (MLP and ART1 networks), we exploit various feature sets calculated from the contour of image; the Hu moments and features obtained with sliding windows. The implementation results on IFN/ENIT database show a high degree of accuracy by applying the majority vote method.