Temporal Multiple Correspondence Analysis for Big Data Mining in Soccer Videos

Yimin Yang, Shu Ching Chen, Mei Ling Shyu

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

3 Scopus citations

Abstract

A multimedia big data mining framework consisting of two phases for interesting event detection in soccer videos has been proposed in this paper. In the pre-processing phase, it utilizes the multi-modal multi-filtering content analysis techniques for shot boundary detection and feature extraction. A pre-filtering process based on domain knowledge analysis is then applied to clean the noise and obtain a candidate set. In the event detection phase, a temporal multiple correspondence analysis (TMCA) algorithm that adopts an indicator weighting scheme is proposed to efficiently and effectively incorporate the temporal semantic information for improving the detection results. Furthermore, another enhanced MCA (EN-MCA) approach is presented to better capture the correspondence between feature items and classes by thoroughly utilizing the pair-wise principal components. Finally, a re-ranking procedure is performed to retrieve the missed interesting event. Our proposed semantic re-ranking framework is evaluated on a large collection of soccer videos for interesting event detection. The experimental results demonstrate the effectiveness of the proposed framework.

Original languageEnglish (US)
Title of host publicationProceedings - 2015 IEEE International Conference on Multimedia Big Data, BigMM 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages64-71
Number of pages8
ISBN (Electronic)9781479986880
DOIs
StatePublished - Jul 9 2015
Event1st IEEE International Conference on Multimedia Big Data, BigMM 2015 - Beijing, China
Duration: Apr 20 2015Apr 22 2015

Publication series

NameProceedings - 2015 IEEE International Conference on Multimedia Big Data, BigMM 2015

Other

Other1st IEEE International Conference on Multimedia Big Data, BigMM 2015
CountryChina
CityBeijing
Period4/20/154/22/15

Keywords

  • Big data
  • event detection
  • multimedia big data mining
  • re-ranking
  • temporal MCA

ASJC Scopus subject areas

  • Information Systems
  • Media Technology

Fingerprint Dive into the research topics of 'Temporal Multiple Correspondence Analysis for Big Data Mining in Soccer Videos'. Together they form a unique fingerprint.

  • Cite this

    Yang, Y., Chen, S. C., & Shyu, M. L. (2015). Temporal Multiple Correspondence Analysis for Big Data Mining in Soccer Videos. In Proceedings - 2015 IEEE International Conference on Multimedia Big Data, BigMM 2015 (pp. 64-71). [7153857] (Proceedings - 2015 IEEE International Conference on Multimedia Big Data, BigMM 2015). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/BigMM.2015.88