Mining high-level features from video using associations and correlations

Lin Lin, Mei-Ling Shyu

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

3 Citations (Scopus)

Abstract

Association rule mining (ARM) has been studied in the areas of content-based multimedia retrieval and semantic concept detection due to its high efficiency and accuracy. Two important processes in mining the association rules for classification are rule generation and rule selection. In this paper, a novel high-level feature detection framework using the ARM technique together with the correlations among the feature-value pairs is proposed. A new association rule mining (ARM) algorithm has been developed, where the N-feature-value pairs are generated using a combined measure based on (1) the existence of the (N-1)-feature-value pairs (where N is larger than 1), (2) the correlation between different N-feature-value pairs and the concept classes through Multiple Correspondence Analysis (MCA), and (3) the similarity representing the harmonic mean of the inter-similarity and intrasimilarity. The final association classification rules are selected by using the calculated harmonic mean of the similarity values. The proposed framework enables the automatic discovery and generation of the N-feature-value pair association rules from the 1-feature-value pairs for classification. Experimenting with 15 high-level features (concepts) and benchmark data sets from TRECVID, our proposed framework achieves promising performance and outperforms three other well-known classifiers (Decision Trees, Support Vector Machine, and Neural Networks) which are commonly used for performance comparison in the TRECVID community.

Original languageEnglish
Title of host publicationICSC 2009 - 2009 IEEE International Conference on Semantic Computing
Pages137-144
Number of pages8
DOIs
StatePublished - Dec 1 2009
EventICSC 2009 - 2009 IEEE International Conference on Semantic Computing - Berkeley, CA, United States
Duration: Sep 14 2009Sep 16 2009

Other

OtherICSC 2009 - 2009 IEEE International Conference on Semantic Computing
CountryUnited States
CityBerkeley, CA
Period9/14/099/16/09

Fingerprint

Association rules
Decision trees
Support vector machines
Classifiers
Semantics
Neural networks

Keywords

  • Association rule mining
  • Concept detection
  • Multiple Correspondence Analysis (MCA)

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Software
  • Electrical and Electronic Engineering

Cite this

Lin, L., & Shyu, M-L. (2009). Mining high-level features from video using associations and correlations. In ICSC 2009 - 2009 IEEE International Conference on Semantic Computing (pp. 137-144). [5298603] https://doi.org/10.1109/ICSC.2009.59

Mining high-level features from video using associations and correlations. / Lin, Lin; Shyu, Mei-Ling.

ICSC 2009 - 2009 IEEE International Conference on Semantic Computing. 2009. p. 137-144 5298603.

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

Lin, L & Shyu, M-L 2009, Mining high-level features from video using associations and correlations. in ICSC 2009 - 2009 IEEE International Conference on Semantic Computing., 5298603, pp. 137-144, ICSC 2009 - 2009 IEEE International Conference on Semantic Computing, Berkeley, CA, United States, 9/14/09. https://doi.org/10.1109/ICSC.2009.59
Lin L, Shyu M-L. Mining high-level features from video using associations and correlations. In ICSC 2009 - 2009 IEEE International Conference on Semantic Computing. 2009. p. 137-144. 5298603 https://doi.org/10.1109/ICSC.2009.59
Lin, Lin ; Shyu, Mei-Ling. / Mining high-level features from video using associations and correlations. ICSC 2009 - 2009 IEEE International Conference on Semantic Computing. 2009. pp. 137-144
@inproceedings{69a0b9a4e2244a56a64e3a75e84faa00,
title = "Mining high-level features from video using associations and correlations",
abstract = "Association rule mining (ARM) has been studied in the areas of content-based multimedia retrieval and semantic concept detection due to its high efficiency and accuracy. Two important processes in mining the association rules for classification are rule generation and rule selection. In this paper, a novel high-level feature detection framework using the ARM technique together with the correlations among the feature-value pairs is proposed. A new association rule mining (ARM) algorithm has been developed, where the N-feature-value pairs are generated using a combined measure based on (1) the existence of the (N-1)-feature-value pairs (where N is larger than 1), (2) the correlation between different N-feature-value pairs and the concept classes through Multiple Correspondence Analysis (MCA), and (3) the similarity representing the harmonic mean of the inter-similarity and intrasimilarity. The final association classification rules are selected by using the calculated harmonic mean of the similarity values. The proposed framework enables the automatic discovery and generation of the N-feature-value pair association rules from the 1-feature-value pairs for classification. Experimenting with 15 high-level features (concepts) and benchmark data sets from TRECVID, our proposed framework achieves promising performance and outperforms three other well-known classifiers (Decision Trees, Support Vector Machine, and Neural Networks) which are commonly used for performance comparison in the TRECVID community.",
keywords = "Association rule mining, Concept detection, Multiple Correspondence Analysis (MCA)",
author = "Lin Lin and Mei-Ling Shyu",
year = "2009",
month = "12",
day = "1",
doi = "10.1109/ICSC.2009.59",
language = "English",
isbn = "9780769538006",
pages = "137--144",
booktitle = "ICSC 2009 - 2009 IEEE International Conference on Semantic Computing",

}

TY - GEN

T1 - Mining high-level features from video using associations and correlations

AU - Lin, Lin

AU - Shyu, Mei-Ling

PY - 2009/12/1

Y1 - 2009/12/1

N2 - Association rule mining (ARM) has been studied in the areas of content-based multimedia retrieval and semantic concept detection due to its high efficiency and accuracy. Two important processes in mining the association rules for classification are rule generation and rule selection. In this paper, a novel high-level feature detection framework using the ARM technique together with the correlations among the feature-value pairs is proposed. A new association rule mining (ARM) algorithm has been developed, where the N-feature-value pairs are generated using a combined measure based on (1) the existence of the (N-1)-feature-value pairs (where N is larger than 1), (2) the correlation between different N-feature-value pairs and the concept classes through Multiple Correspondence Analysis (MCA), and (3) the similarity representing the harmonic mean of the inter-similarity and intrasimilarity. The final association classification rules are selected by using the calculated harmonic mean of the similarity values. The proposed framework enables the automatic discovery and generation of the N-feature-value pair association rules from the 1-feature-value pairs for classification. Experimenting with 15 high-level features (concepts) and benchmark data sets from TRECVID, our proposed framework achieves promising performance and outperforms three other well-known classifiers (Decision Trees, Support Vector Machine, and Neural Networks) which are commonly used for performance comparison in the TRECVID community.

AB - Association rule mining (ARM) has been studied in the areas of content-based multimedia retrieval and semantic concept detection due to its high efficiency and accuracy. Two important processes in mining the association rules for classification are rule generation and rule selection. In this paper, a novel high-level feature detection framework using the ARM technique together with the correlations among the feature-value pairs is proposed. A new association rule mining (ARM) algorithm has been developed, where the N-feature-value pairs are generated using a combined measure based on (1) the existence of the (N-1)-feature-value pairs (where N is larger than 1), (2) the correlation between different N-feature-value pairs and the concept classes through Multiple Correspondence Analysis (MCA), and (3) the similarity representing the harmonic mean of the inter-similarity and intrasimilarity. The final association classification rules are selected by using the calculated harmonic mean of the similarity values. The proposed framework enables the automatic discovery and generation of the N-feature-value pair association rules from the 1-feature-value pairs for classification. Experimenting with 15 high-level features (concepts) and benchmark data sets from TRECVID, our proposed framework achieves promising performance and outperforms three other well-known classifiers (Decision Trees, Support Vector Machine, and Neural Networks) which are commonly used for performance comparison in the TRECVID community.

KW - Association rule mining

KW - Concept detection

KW - Multiple Correspondence Analysis (MCA)

UR - http://www.scopus.com/inward/record.url?scp=73449130148&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=73449130148&partnerID=8YFLogxK

U2 - 10.1109/ICSC.2009.59

DO - 10.1109/ICSC.2009.59

M3 - Conference contribution

AN - SCOPUS:73449130148

SN - 9780769538006

SP - 137

EP - 144

BT - ICSC 2009 - 2009 IEEE International Conference on Semantic Computing

ER -