Robust Scene Extraction Using Multi-Stream HMMs for Baseball Broadcast

Nguyen Huu BACH; Koichi SHINODA; Sadaoki FURUI

doi:10.1093/ietisy/e89-d.9.2553

Robust Scene Extraction Using Multi-Stream HMMs for Baseball Broadcast

Nguyen Huu BACH, Koichi SHINODA, Sadaoki FURUI

Full Text Views

0

Share
Cite this

Summary :

In this paper, we propose a robust statistical framework for extracting scenes from a baseball broadcast video. We apply multi-stream hidden Markov models (HMMs) to control the weights among different features. To achieve a large robustness against new scenes, we used a common simple structure for all the HMMs. In addition, scene segmentation and unsupervised adaptation were applied to achieve greater robustness against differences in environmental conditions among games. The F-measure of scene-extracting experiments for eight types of scene from 4.5 hours of digest data was 77.4% and was increased to 78.7% by applying scene segmentation. Furthermore, the unsupervised adaptation method improved precision by 2.7 points to 81.4%. These results confirm the effectiveness of our framework.

Publication: IEICE TRANSACTIONS on Information Vol.E89-D No.9 pp.2553-2561

Publication Date: 2006/09/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1093/ietisy/e89-d.9.2553

Type of Manuscript: PAPER

Category: Image Processing and Video Processing

Cite this

Copy

Nguyen Huu BACH, Koichi SHINODA, Sadaoki FURUI, "Robust Scene Extraction Using Multi-Stream HMMs for Baseball Broadcast" in IEICE TRANSACTIONS on Information, vol. E89-D, no. 9, pp. 2553-2561, September 2006, doi: 10.1093/ietisy/e89-d.9.2553.
Abstract: In this paper, we propose a robust statistical framework for extracting scenes from a baseball broadcast video. We apply multi-stream hidden Markov models (HMMs) to control the weights among different features. To achieve a large robustness against new scenes, we used a common simple structure for all the HMMs. In addition, scene segmentation and unsupervised adaptation were applied to achieve greater robustness against differences in environmental conditions among games. The F-measure of scene-extracting experiments for eight types of scene from 4.5 hours of digest data was 77.4% and was increased to 78.7% by applying scene segmentation. Furthermore, the unsupervised adaptation method improved precision by 2.7 points to 81.4%. These results confirm the effectiveness of our framework.
URL: https://globals.ieice.org/en_transactions/information/10.1093/ietisy/e89-d.9.2553/_p

Copy

@ARTICLE{e89-d_9_2553,
author={Nguyen Huu BACH, Koichi SHINODA, Sadaoki FURUI, },
journal={IEICE TRANSACTIONS on Information},
title={Robust Scene Extraction Using Multi-Stream HMMs for Baseball Broadcast},
year={2006},
volume={E89-D},
number={9},
pages={2553-2561},
abstract={In this paper, we propose a robust statistical framework for extracting scenes from a baseball broadcast video. We apply multi-stream hidden Markov models (HMMs) to control the weights among different features. To achieve a large robustness against new scenes, we used a common simple structure for all the HMMs. In addition, scene segmentation and unsupervised adaptation were applied to achieve greater robustness against differences in environmental conditions among games. The F-measure of scene-extracting experiments for eight types of scene from 4.5 hours of digest data was 77.4% and was increased to 78.7% by applying scene segmentation. Furthermore, the unsupervised adaptation method improved precision by 2.7 points to 81.4%. These results confirm the effectiveness of our framework.},
keywords={},
doi={10.1093/ietisy/e89-d.9.2553},
ISSN={1745-1361},
month={September},}

Copy

TY - JOUR
TI - Robust Scene Extraction Using Multi-Stream HMMs for Baseball Broadcast
T2 - IEICE TRANSACTIONS on Information
SP - 2553
EP - 2561
AU - Nguyen Huu BACH
AU - Koichi SHINODA
AU - Sadaoki FURUI
PY - 2006
DO - 10.1093/ietisy/e89-d.9.2553
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E89-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2006
AB - In this paper, we propose a robust statistical framework for extracting scenes from a baseball broadcast video. We apply multi-stream hidden Markov models (HMMs) to control the weights among different features. To achieve a large robustness against new scenes, we used a common simple structure for all the HMMs. In addition, scene segmentation and unsupervised adaptation were applied to achieve greater robustness against differences in environmental conditions among games. The F-measure of scene-extracting experiments for eight types of scene from 4.5 hours of digest data was 77.4% and was increased to 78.7% by applying scene segmentation. Furthermore, the unsupervised adaptation method improved precision by 2.7 points to 81.4%. These results confirm the effectiveness of our framework.
ER -