Multiple models of Bayesian networks applied to offline recognition of Arabic handwritten city names
This addresses the problem of multi-writer Arabic handwriting recognition for domain-specific applications, but it is incremental as it applies existing Bayesian network methods to a new dataset.
The paper tackles offline recognition of Arabic handwritten city names by proposing an approach using multiple Bayesian networks, achieving good recognition rates with FAN and DBN outperforming other variants.
In this paper we address the problem of offline Arabic handwriting word recognition. Off-line recognition of handwritten words is a difficult task due to the high variability and uncertainty of human writing. The majority of the recent systems are constrained by the size of the lexicon to deal with and the number of writers. In this paper, we propose an approach for multi-writers Arabic handwritten words recognition using multiple Bayesian networks. First, we cut the image in several blocks. For each block, we compute a vector of descriptors. Then, we use K-means to cluster the low-level features including Zernik and Hu moments. Finally, we apply four variants of Bayesian networks classifiers (Naïve Bayes, Tree Augmented Naïve Bayes (TAN), Forest Augmented Naïve Bayes (FAN) and DBN (dynamic bayesian network) to classify the whole image of tunisian city name. The results demonstrate FAN and DBN outperform good recognition rates