A Novel HMM Decoding Algorithm Permitting Long-Term Dependencies and Its Application to Handwritten Word Recognition

Volkmar Frinken, Ryosuke Kakisako, Seiichi Uchida

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

A new decoding for hidden Markov models is presented. As opposed to the commonly used Viterbi algorithm, it is based on the Min-Cut/Max-Flow algorithm instead of dynamic programming. Therefore non-Markovian long-term dependencies can easily be added to influence the decoding path while still finding the optimal decoding in polynomial time. We demonstrate through an experimental evaluation how these constraints can be used to improve an HMM-based handwritten word recognition system that model words via linear character-HMM by restricting the length of each character.

Original languageEnglish
Title of host publicationProceedings - 14th International Conference on Frontiers in Handwriting Recognition, ICFHR 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages128-133
Number of pages6
ISBN (Electronic)9781479943340
DOIs
Publication statusPublished - Dec 9 2014
Event14th International Conference on Frontiers in Handwriting Recognition, ICFHR 2014 - Hersonissos, Crete Island, Greece
Duration: Sep 1 2014Sep 4 2014

Publication series

NameProceedings of International Conference on Frontiers in Handwriting Recognition, ICFHR
Volume2014-December
ISSN (Print)2167-6445
ISSN (Electronic)2167-6453

Other

Other14th International Conference on Frontiers in Handwriting Recognition, ICFHR 2014
CountryGreece
CityHersonissos, Crete Island
Period9/1/149/4/14

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Computer Vision and Pattern Recognition

Cite this

Frinken, V., Kakisako, R., & Uchida, S. (2014). A Novel HMM Decoding Algorithm Permitting Long-Term Dependencies and Its Application to Handwritten Word Recognition. In Proceedings - 14th International Conference on Frontiers in Handwriting Recognition, ICFHR 2014 (pp. 128-133). [6981008] (Proceedings of International Conference on Frontiers in Handwriting Recognition, ICFHR; Vol. 2014-December). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICFHR.2014.29