Early recognition of sequential patterns by classifier combination

Seiichi Uchida, Kazuma Amamoto

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

12 Citations (Scopus)

Abstract

This paper proposes an early recognition method, i.e., a method for recognizing sequential patterns at their beginning parts. The method is based on a combination of frame classifiers prepared at individual frames. The training patterns misrecognized by the frame classifier at a certain frame are heavily weighted for the complementary training of the frame classifier at the next frame. The method was applied to an online character recognition task for showing its usefulness.

Original languageEnglish
Title of host publication2008 19th International Conference on Pattern Recognition, ICPR 2008
Publication statusPublished - 2008
Event2008 19th International Conference on Pattern Recognition, ICPR 2008 - Tampa, FL, United States
Duration: Dec 8 2008Dec 11 2008

Other

Other2008 19th International Conference on Pattern Recognition, ICPR 2008
CountryUnited States
CityTampa, FL
Period12/8/0812/11/08

    Fingerprint

All Science Journal Classification (ASJC) codes

  • Computer Vision and Pattern Recognition

Cite this

Uchida, S., & Amamoto, K. (2008). Early recognition of sequential patterns by classifier combination. In 2008 19th International Conference on Pattern Recognition, ICPR 2008 [4761137]