### 抄録

In this paper, an action recognition system was invented by proposing a compact 3D descriptor to represent action information, and employing self-organizing map (SOM) to learn and recognize actions. Histogram Of Gradient 3D (HOG3D) performed better among currently used descriptors for action recognition. However, the calculation of the descriptor is quite complex. Furthermore, it used a vector with 960 elements to describe one interest point. Therefore, we proposed a compact descriptor, which shortened the support region of interest points, combined symmetric bins after orientation quantization. In addition, the top value bin of quantized vector was kept instead of setting threshold experimentally. Comparing with HOG3D, our descriptor used 80 bins to describe a point, which reduced much computation complexity. The compact descriptor was used to learn and recognize actions considering the probability of local features in SOM, and the results showed that our system outperformed others both on KTH and Hollywood datasets.

元の言語 | 英語 |
---|---|

ホスト出版物のタイトル | Neural Information Processing |

ホスト出版物のサブタイトル | Models and Applications - 17th International Conference, ICONIP 2010, Proceedings |

ページ | 391-398 |

ページ数 | 8 |

エディション | PART 2 |

DOI | |

出版物ステータス | 出版済み - 12 21 2010 |

イベント | 17th International Conference on Neural Information Processing, ICONIP 2010 - Sydney, NSW, オーストラリア 継続期間: 11 22 2010 → 11 25 2010 |

### 出版物シリーズ

名前 | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|

番号 | PART 2 |

巻 | 6444 LNCS |

ISSN（印刷物） | 0302-9743 |

ISSN（電子版） | 1611-3349 |

### その他

その他 | 17th International Conference on Neural Information Processing, ICONIP 2010 |
---|---|

国 | オーストラリア |

市 | Sydney, NSW |

期間 | 11/22/10 → 11/25/10 |

### Fingerprint

### All Science Journal Classification (ASJC) codes

- Theoretical Computer Science
- Computer Science(all)

### これを引用

*Neural Information Processing: Models and Applications - 17th International Conference, ICONIP 2010, Proceedings*(PART 2 版, pp. 391-398). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 巻数 6444 LNCS, 番号 PART 2). https://doi.org/10.1007/978-3-642-17534-3_48

**Human action recognition by SOM considering the probability of spatio-temporal features.** / Ji, Yanli; Shimada, Atsushi; Taniguchi, Rin Ichiro.

研究成果: 著書/レポートタイプへの貢献 › 会議での発言

*Neural Information Processing: Models and Applications - 17th International Conference, ICONIP 2010, Proceedings.*PART 2 Edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 番号 PART 2, 巻. 6444 LNCS, pp. 391-398, 17th International Conference on Neural Information Processing, ICONIP 2010, Sydney, NSW, オーストラリア, 11/22/10. https://doi.org/10.1007/978-3-642-17534-3_48

}

TY - GEN

T1 - Human action recognition by SOM considering the probability of spatio-temporal features

AU - Ji, Yanli

AU - Shimada, Atsushi

AU - Taniguchi, Rin Ichiro

PY - 2010/12/21

Y1 - 2010/12/21

N2 - In this paper, an action recognition system was invented by proposing a compact 3D descriptor to represent action information, and employing self-organizing map (SOM) to learn and recognize actions. Histogram Of Gradient 3D (HOG3D) performed better among currently used descriptors for action recognition. However, the calculation of the descriptor is quite complex. Furthermore, it used a vector with 960 elements to describe one interest point. Therefore, we proposed a compact descriptor, which shortened the support region of interest points, combined symmetric bins after orientation quantization. In addition, the top value bin of quantized vector was kept instead of setting threshold experimentally. Comparing with HOG3D, our descriptor used 80 bins to describe a point, which reduced much computation complexity. The compact descriptor was used to learn and recognize actions considering the probability of local features in SOM, and the results showed that our system outperformed others both on KTH and Hollywood datasets.

AB - In this paper, an action recognition system was invented by proposing a compact 3D descriptor to represent action information, and employing self-organizing map (SOM) to learn and recognize actions. Histogram Of Gradient 3D (HOG3D) performed better among currently used descriptors for action recognition. However, the calculation of the descriptor is quite complex. Furthermore, it used a vector with 960 elements to describe one interest point. Therefore, we proposed a compact descriptor, which shortened the support region of interest points, combined symmetric bins after orientation quantization. In addition, the top value bin of quantized vector was kept instead of setting threshold experimentally. Comparing with HOG3D, our descriptor used 80 bins to describe a point, which reduced much computation complexity. The compact descriptor was used to learn and recognize actions considering the probability of local features in SOM, and the results showed that our system outperformed others both on KTH and Hollywood datasets.

UR - http://www.scopus.com/inward/record.url?scp=78650192992&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78650192992&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-17534-3_48

DO - 10.1007/978-3-642-17534-3_48

M3 - Conference contribution

AN - SCOPUS:78650192992

SN - 3642175333

SN - 9783642175336

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 391

EP - 398

BT - Neural Information Processing

ER -