DMCA
Exploring the Use of Conditional Random Field Models and HMMs for Historical Handwritten Document Recognition
| Venue: | the Proceedings of the 2nd IEEE International Conference on Document Image Analysis for Libraries (DIAL |
| Citations: | 13 - 4 self |
BibTeX
@INPROCEEDINGS{Feng_exploringthe,
author = {Shaolei Feng and R. Manmatha and Andrew Mccallum},
title = {Exploring the Use of Conditional Random Field Models and HMMs for Historical Handwritten Document Recognition},
booktitle = {the Proceedings of the 2nd IEEE International Conference on Document Image Analysis for Libraries (DIAL},
year = {},
pages = {30--37}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
In this paper we explore different approaches for improving the performance of dependency models on discrete features for handwriting recognition. Hidden Markov Models have often been used for handwriting recognition. Conditional random fields (CRF’s) allow for more general dependencies and we investigate their use. We believe that this is the first attempt at apply CRF’s for handwriting recognition. We show that on the whole word recognition task, the CRF performs better than a HMM on a publicly available standard dataset of 20 pages of George Washington’s manuscripts. The scale space for the whole word recognition task is large- almost 1200 states. To make CRF computation tractable we use beam search to make inference more efficient using three different approaches. Better improvement can be obtained using the HMM by directly smoothing the discrete features using the collection frequencies. This shows the importance of smoothing and also indicates the difficulty of training CRF’s when large state spaces are involved. 1
Keyphrases
conditional random field model historical handwritten document recognition different approach whole word recognition task dependency model apply crf conditional random field crf performs scale space general dependency collection frequency beam search hidden markov model discrete feature george washington first attempt large state space available standard dataset crf computation tractable
