Off-line recognition handwriting Arabic words using combination of multiple classifiers

We present in this paper a system of Arabic handwriting recognition based on combining methods of decision fusion approach. The proposed approach introduces a methodology using the HMM-Toolkit (HTK) for a rapid implementation of our designed recognition system. After the image preprocessing, the text is segmented into lines, the obtained images are then used for features extraction with Sliding window technique. These features are extracted on binary images of characters and are modeled separately using Hidden Markov Models classifiers. The combination of the multiple HMMs classifiers was applied by using the different methods of decision fusion approach. The proposed system is evaluated using the IFN/ENIT database. Experimental results for Arabic handwritten recognition demonstrate that the Weighted Majority Voting (WMV) combination method have given better recognition rate 76.54% in top1, with Gaussian distribution.