Robust Markerless Tracking of Knee Joint for Indoor and Outdoor Cycling

• Item type: SIG Technical Reports(1)
• 公開日: 2019-02-28
• タイトル: Robust Markerless Tracking of Knee Joint for Indoor and Outdoor Cycling
• タイトル / 言語: en
• タイトル: Robust Markerless Tracking of Knee Joint for Indoor and Outdoor Cycling
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: テーマセッション
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: Nara Institute of Science and Technology
• 著者所属: Kyoto University Hospital
• 著者所属: Nara Institute of Science and Technology
• 著者所属: Nara Institute of Science and Technology
• 著者所属: Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Kyoto University Hospital
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者所属(英): en
• 著者所属(英): Nara Institute of Science and Technology
• 著者名: Oral, Kaplan; Goshiro, Yamamoto; Takafumi, Taketomi; Alexander, Plopski; Hirokazu, Kato
• 著者名(英): Oral, Kaplan; Goshiro, Yamamoto; Takafumi, Taketomi; Alexander, Plopski; Hirokazu, Kato
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: In this work, we demonstrate the applicability of an existing deep learning based toolbox to indoor-outdoor cycling for realizing robust markerless tracking of knee joint. Cycling has become a ubiquitous physical activity worldwide for recreation, commuting, or sport. It is a low-impact non-weight bearing form of physical activity due to body weight being carried by a bicycle. However, this does not guarantee an experience free from injuries. Besides falls, the repetitive nature of cycling and monotonous loading of joints are consistently associated with overuse injuries. Among all, knee overuse remains an ill-defined injury type with anecdotal treatment approaches. Although biomedical research consider numerous factors as originators, research efforts utilize quantitative data captured through stationary indoor scenarios alone due to technological limitations. Therefore, we consider a two-part video-based framework for cycling to enable indoor-outdoor tracking of knee movement and trajectory visualizations. In this paper, we focus on former and describe our preliminary studies on tracking to demonstrate its applicability to cycling. Furthermore, we clarify the place of our work in literature by introducing the ongoing research, and formulate several future directions that may provide new insights into knee overuse injuries. We consider our approach promising for realizing cost-efficient monitoring of knee joint during indoor-outdoor cycling.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: In this work, we demonstrate the applicability of an existing deep learning based toolbox to indoor-outdoor cycling for realizing robust markerless tracking of knee joint. Cycling has become a ubiquitous physical activity worldwide for recreation, commuting, or sport. It is a low-impact non-weight bearing form of physical activity due to body weight being carried by a bicycle. However, this does not guarantee an experience free from injuries. Besides falls, the repetitive nature of cycling and monotonous loading of joints are consistently associated with overuse injuries. Among all, knee overuse remains an ill-defined injury type with anecdotal treatment approaches. Although biomedical research consider numerous factors as originators, research efforts utilize quantitative data captured through stationary indoor scenarios alone due to technological limitations. Therefore, we consider a two-part video-based framework for cycling to enable indoor-outdoor tracking of knee movement and trajectory visualizations. In this paper, we focus on former and describe our preliminary studies on tracking to demonstrate its applicability to cycling. Furthermore, we clarify the place of our work in literature by introducing the ongoing research, and formulate several future directions that may provide new insights into knee overuse injuries. We consider our approach promising for realizing cost-efficient monitoring of knee joint during indoor-outdoor cycling.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AA11131797
• 書誌情報: 研究報告コンピュータビジョンとイメージメディア（CVIM）; 巻 2019-CVIM-216, 号 7, p. 1-4, 発行日 2019-02-28
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2188-8701
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会