Comparison of global and cascading recognition systems applied to multi-font arabic text
Proceedings of the 10th ACM symposium on Document engineering
Pattern Recognition Letters
Offline arabic handwritten text recognition: A Survey
ACM Computing Surveys (CSUR)
Hi-index | 0.00 |
We present in this paper a system for recognition ofprinted Arabic text based on Hidden Markov Models(HMM). While HMMs have been successfully used inthe past for such a task, we report here on significantimprovements of the recognition performance with theintroduction of minimum and maximum durationmodels. The improvements allow us to build a systemworking in open vocabulary mode, i.e., without anylimitations on the size of the vocabulary. Theevaluation of our system is performed using HTK(Hidden Markov Model Toolkit) on a database of wordimages that are synthetically generated.