Speech-annotated photo retrieval using syllable-transformed patterns

Chung-Hsien Wu, Chien Lin Huang, Wei Chuan Lee, Yu Sheng Lai

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

This study presents a novel indexing and retrieval scheme for digital photos with speech annotations based on syllable-transformed image-like patterns. Speech recognition error and out-of-vocabulary (OOV) problems generally result in incorrect indexing and degrade the retrieval performance. In this study, the recognized n-best candidates used to deal with recognition error problems are transformed into an image-like pattern using multidimensional scaling. A hybrid mechanism integrating syllables, characters, words, and image-like patterns is exploited for speech indexing and retrieval. Experiments show the hybrid indexing method integrating the syllable-transformed image-like patterns can achieve a better result compared to previous indexing methods.

Original languageEnglish
Pages (from-to)6-9
Number of pages4
JournalIEEE Signal Processing Letters
Volume16
Issue number1
DOIs
Publication statusPublished - 2009 Jan 5

Fingerprint

Indexing
Retrieval
Speech recognition
Speech Recognition
Experiments
Annotation
Speech
Scaling
Experiment

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Electrical and Electronic Engineering
  • Applied Mathematics

Cite this

Wu, Chung-Hsien ; Huang, Chien Lin ; Lee, Wei Chuan ; Lai, Yu Sheng. / Speech-annotated photo retrieval using syllable-transformed patterns. In: IEEE Signal Processing Letters. 2009 ; Vol. 16, No. 1. pp. 6-9.
@article{a272b640a0f14a0e8e3448723a5993ec,
title = "Speech-annotated photo retrieval using syllable-transformed patterns",
abstract = "This study presents a novel indexing and retrieval scheme for digital photos with speech annotations based on syllable-transformed image-like patterns. Speech recognition error and out-of-vocabulary (OOV) problems generally result in incorrect indexing and degrade the retrieval performance. In this study, the recognized n-best candidates used to deal with recognition error problems are transformed into an image-like pattern using multidimensional scaling. A hybrid mechanism integrating syllables, characters, words, and image-like patterns is exploited for speech indexing and retrieval. Experiments show the hybrid indexing method integrating the syllable-transformed image-like patterns can achieve a better result compared to previous indexing methods.",
author = "Chung-Hsien Wu and Huang, {Chien Lin} and Lee, {Wei Chuan} and Lai, {Yu Sheng}",
year = "2009",
month = "1",
day = "5",
doi = "10.1109/LSP.2008.2008490",
language = "English",
volume = "16",
pages = "6--9",
journal = "IEEE Signal Processing Letters",
issn = "1070-9908",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "1",

}

Speech-annotated photo retrieval using syllable-transformed patterns. / Wu, Chung-Hsien; Huang, Chien Lin; Lee, Wei Chuan; Lai, Yu Sheng.

In: IEEE Signal Processing Letters, Vol. 16, No. 1, 05.01.2009, p. 6-9.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Speech-annotated photo retrieval using syllable-transformed patterns

AU - Wu, Chung-Hsien

AU - Huang, Chien Lin

AU - Lee, Wei Chuan

AU - Lai, Yu Sheng

PY - 2009/1/5

Y1 - 2009/1/5

N2 - This study presents a novel indexing and retrieval scheme for digital photos with speech annotations based on syllable-transformed image-like patterns. Speech recognition error and out-of-vocabulary (OOV) problems generally result in incorrect indexing and degrade the retrieval performance. In this study, the recognized n-best candidates used to deal with recognition error problems are transformed into an image-like pattern using multidimensional scaling. A hybrid mechanism integrating syllables, characters, words, and image-like patterns is exploited for speech indexing and retrieval. Experiments show the hybrid indexing method integrating the syllable-transformed image-like patterns can achieve a better result compared to previous indexing methods.

AB - This study presents a novel indexing and retrieval scheme for digital photos with speech annotations based on syllable-transformed image-like patterns. Speech recognition error and out-of-vocabulary (OOV) problems generally result in incorrect indexing and degrade the retrieval performance. In this study, the recognized n-best candidates used to deal with recognition error problems are transformed into an image-like pattern using multidimensional scaling. A hybrid mechanism integrating syllables, characters, words, and image-like patterns is exploited for speech indexing and retrieval. Experiments show the hybrid indexing method integrating the syllable-transformed image-like patterns can achieve a better result compared to previous indexing methods.

UR - http://www.scopus.com/inward/record.url?scp=58049207761&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=58049207761&partnerID=8YFLogxK

U2 - 10.1109/LSP.2008.2008490

DO - 10.1109/LSP.2008.2008490

M3 - Article

VL - 16

SP - 6

EP - 9

JO - IEEE Signal Processing Letters

JF - IEEE Signal Processing Letters

SN - 1070-9908

IS - 1

ER -