Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs

Shengchang Chen; Shujing Lu; Ying Wen; Yue Lu

doi:10.1109/ICDAR.2015.7333742

Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs

Shengchang Chen, Shujing Lu, Ying Wen, Yue Lu

School of Communication and Electronic Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

Different recognizers may result in different mistakes when they are used to recognize a Chinese address. In this paper, we present a method of combining multiple Chinese address recognition outputs to improve Chinese address recognition accuracy. The method first employs multiple sequence alignment to generate a lattice of candidate hypotheses from multiple different recognizer outputs and then applies statistical language model to choose the maximum likelihood candidate sequence. Taking the maximum as the final decision, the performance of our method is superior, compared to the single recognizers and Miyao's method. The experiments on the address images of real envelopes demonstrate that the proposed method increases the character recognition accuracy rate from 95.80% to 98.38%, with 61.30% error reduction. Furthermore, the corrected sorting rate of an automatic mail sorting system increases from 84.11% to 93.72%.

Original language	English
Title of host publication	13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings
Publisher	IEEE Computer Society
Pages	151-155
Number of pages	5
ISBN (Electronic)	9781479918058
DOIs	https://doi.org/10.1109/ICDAR.2015.7333742
State	Published - 20 Nov 2015
Event	13th International Conference on Document Analysis and Recognition, ICDAR 2015 - Nancy, France Duration: 23 Aug 2015 → 26 Aug 2015

Publication series

Name	Proceedings of the International Conference on Document Analysis and Recognition, ICDAR
Volume	2015-November
ISSN (Print)	1520-5363

Conference

Conference	13th International Conference on Document Analysis and Recognition, ICDAR 2015
Country/Territory	France
City	Nancy
Period	23/08/15 → 26/08/15

Keywords

minimum edit distance
multiple Chinese address recognition outputs
multiple sequence alignment
statistical language model

Access to Document

10.1109/ICDAR.2015.7333742

Cite this

Chen, S., Lu, S., Wen, Y., & Lu, Y. (2015). Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs. In 13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings (pp. 151-155). Article 7333742 (Proceedings of the International Conference on Document Analysis and Recognition, ICDAR; Vol. 2015-November). IEEE Computer Society. https://doi.org/10.1109/ICDAR.2015.7333742

Chen, Shengchang ; Lu, Shujing ; Wen, Ying et al. / Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs. 13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings. IEEE Computer Society, 2015. pp. 151-155 (Proceedings of the International Conference on Document Analysis and Recognition, ICDAR).

@inproceedings{8f957893d7934e9cab778b2d33368bca,

title = "Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs",

abstract = "Different recognizers may result in different mistakes when they are used to recognize a Chinese address. In this paper, we present a method of combining multiple Chinese address recognition outputs to improve Chinese address recognition accuracy. The method first employs multiple sequence alignment to generate a lattice of candidate hypotheses from multiple different recognizer outputs and then applies statistical language model to choose the maximum likelihood candidate sequence. Taking the maximum as the final decision, the performance of our method is superior, compared to the single recognizers and Miyao's method. The experiments on the address images of real envelopes demonstrate that the proposed method increases the character recognition accuracy rate from 95.80\% to 98.38\%, with 61.30\% error reduction. Furthermore, the corrected sorting rate of an automatic mail sorting system increases from 84.11\% to 93.72\%.",

keywords = "minimum edit distance, multiple Chinese address recognition outputs, multiple sequence alignment, statistical language model",

author = "Shengchang Chen and Shujing Lu and Ying Wen and Yue Lu",

note = "Publisher Copyright: {\textcopyright} 2015 IEEE.; 13th International Conference on Document Analysis and Recognition, ICDAR 2015 ; Conference date: 23-08-2015 Through 26-08-2015",

year = "2015",

month = nov,

day = "20",

doi = "10.1109/ICDAR.2015.7333742",

language = "英语",

series = "Proceedings of the International Conference on Document Analysis and Recognition, ICDAR",

publisher = "IEEE Computer Society",

pages = "151--155",

booktitle = "13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings",

address = "美国",

}

Chen, S, Lu, S, Wen, Y & Lu, Y 2015, Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs. in 13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings., 7333742, Proceedings of the International Conference on Document Analysis and Recognition, ICDAR, vol. 2015-November, IEEE Computer Society, pp. 151-155, 13th International Conference on Document Analysis and Recognition, ICDAR 2015, Nancy, France, 23/08/15. https://doi.org/10.1109/ICDAR.2015.7333742

Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs. / Chen, Shengchang; Lu, Shujing; Wen, Ying et al.
13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings. IEEE Computer Society, 2015. p. 151-155 7333742 (Proceedings of the International Conference on Document Analysis and Recognition, ICDAR; Vol. 2015-November).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs

AU - Chen, Shengchang

AU - Lu, Shujing

AU - Wen, Ying

AU - Lu, Yue

PY - 2015/11/20

Y1 - 2015/11/20

N2 - Different recognizers may result in different mistakes when they are used to recognize a Chinese address. In this paper, we present a method of combining multiple Chinese address recognition outputs to improve Chinese address recognition accuracy. The method first employs multiple sequence alignment to generate a lattice of candidate hypotheses from multiple different recognizer outputs and then applies statistical language model to choose the maximum likelihood candidate sequence. Taking the maximum as the final decision, the performance of our method is superior, compared to the single recognizers and Miyao's method. The experiments on the address images of real envelopes demonstrate that the proposed method increases the character recognition accuracy rate from 95.80% to 98.38%, with 61.30% error reduction. Furthermore, the corrected sorting rate of an automatic mail sorting system increases from 84.11% to 93.72%.

AB - Different recognizers may result in different mistakes when they are used to recognize a Chinese address. In this paper, we present a method of combining multiple Chinese address recognition outputs to improve Chinese address recognition accuracy. The method first employs multiple sequence alignment to generate a lattice of candidate hypotheses from multiple different recognizer outputs and then applies statistical language model to choose the maximum likelihood candidate sequence. Taking the maximum as the final decision, the performance of our method is superior, compared to the single recognizers and Miyao's method. The experiments on the address images of real envelopes demonstrate that the proposed method increases the character recognition accuracy rate from 95.80% to 98.38%, with 61.30% error reduction. Furthermore, the corrected sorting rate of an automatic mail sorting system increases from 84.11% to 93.72%.

KW - minimum edit distance

KW - multiple Chinese address recognition outputs

KW - multiple sequence alignment

KW - statistical language model

UR - https://www.scopus.com/pages/publications/84962536306

U2 - 10.1109/ICDAR.2015.7333742

DO - 10.1109/ICDAR.2015.7333742

M3 - 会议稿件

AN - SCOPUS:84962536306

T3 - Proceedings of the International Conference on Document Analysis and Recognition, ICDAR

SP - 151

EP - 155

BT - 13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings

PB - IEEE Computer Society

T2 - 13th International Conference on Document Analysis and Recognition, ICDAR 2015

Y2 - 23 August 2015 through 26 August 2015

ER -

Chen S, Lu S, Wen Y , Lu Y. Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs. In 13th IAPR International Conference on Document Analysis and Recognition, ICDAR 2015 - Conference Proceedings. IEEE Computer Society. 2015. p. 151-155. 7333742. (Proceedings of the International Conference on Document Analysis and Recognition, ICDAR). doi: 10.1109/ICDAR.2015.7333742

Using multiple sequence alignment and statistical language model to integrate multiple Chinese address recognition outputs

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this