المستخلص: |
Our work is one of many researches attempting to solve the problem of automatic vision of handwritten arabic words. We are proposing a method based on psychological studies about human reading methods. These studies suggest that humans focus on words holistically in very short time, and if parts of word becomes incomprehensive, the context of letters or neighbouring words would help. Despite these observations, most methods in handwritten recognition in general go on segmenting the word into graphems in order to recognise letters first, then afterwards recognise words; which is just the contrary of the natural human reading process. We are trying in the present work to create a model that simulates the human model of reading and to produce a system that implements it
|