Offline Recognition of Unconstrained Handwritten Texts Using HMMs and Statistical Language Models
IEEE Transactions on Pattern Analysis and Machine Intelligence
| Hi-index | 0.00 |
The output of handwritten word recognizers (WR) tends to be very noisy due to various factors. In order to compensate for this behaviour, several choices of the WR must be initially considered. In the case of handwritten sentence/phrase recognition, linguistic constraints may be applied in order to improve the results of the WR. This paper discusses two statistical methods of applying linguistic constraints to the output of an WR on input consisting of sentences/phrases. The first is based on collocations and can be used to promote lower ranked word choices or to propose new words. The second is a Markov model of syntax and is based on syntactic categories (tags) associated with words. In each case, we show the improvement in the word recognition rate as a result of applying these constraints.