ML Anthology

Typical Sequences Extraction and Recognition

Gengyu Ma, Xueyin Lin
ECCV 2004 pp. 60-71
doi:10.1007/978-3-540-24837-8_7 /eccv/2004/ma2004eccv-typical/

Abstract

This paper presented a temporal sequence analyzing method, aiming at the extraction of typical sequences from an unlabeled dataset. The extraction procedure is based on HMM training and hierarchical separation of WTOM (Weighted Transition Occurring Matrix). During the extraction, HMMs are built each for a kind of typical sequence. Then Threshold Model is used to segment and recognize continuous sequence. The method has been tested on unsupervised event analysis in video surveillance and model learning of athlete actions.

PDF ECCV Semantic Scholar

Cite

Text

Ma and Lin. "Typical Sequences Extraction and Recognition." European Conference on Computer Vision, 2004. doi:10.1007/978-3-540-24837-8_7

Markdown

[Ma and Lin. "Typical Sequences Extraction and Recognition." European Conference on Computer Vision, 2004.](https://mlanthology.org/eccv/2004/ma2004eccv-typical/) doi:10.1007/978-3-540-24837-8_7

BibTeX

@inproceedings{ma2004eccv-typical,
  title     = {{Typical Sequences Extraction and Recognition}},
  author    = {Ma, Gengyu and Lin, Xueyin},
  booktitle = {European Conference on Computer Vision},
  year      = {2004},
  pages     = {60-71},
  doi       = {10.1007/978-3-540-24837-8_7},
  url       = {https://mlanthology.org/eccv/2004/ma2004eccv-typical/}
}