Preliminary Trial of Activity Recognition of Medical Technicians using Accelerometers

• Item type: SIG Technical Reports(1)
• 公開日: 2022-11-25
• タイトル: Preliminary Trial of Activity Recognition of Medical Technicians using Accelerometers
• タイトル / 言語: en
• タイトル: Preliminary Trial of Activity Recognition of Medical Technicians using Accelerometers
• 言語: eng
• キーワード / 主題Scheme: Other
• キーワード / 主題: 行動認識
• 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: Kyushu Institute of Technology
• 著者所属: Kyushu Institute of Technology
• 著者所属: Kyushu Institute of Technology
• 著者所属: Saiseikai Kumamoto Hospital
• 著者所属: Saiseikai Kumamoto Hospital
• 著者所属: Kyushu University Hospital
• 著者所属: Kyushu Institute of Technology
• 著者所属(英): en
• 著者所属(英): Kyushu Institute of Technology
• 著者所属(英): en
• 著者所属(英): Kyushu Institute of Technology
• 著者所属(英): en
• 著者所属(英): Kyushu Institute of Technology
• 著者所属(英): en
• 著者所属(英): Saiseikai Kumamoto Hospital
• 著者所属(英): en
• 著者所属(英): Saiseikai Kumamoto Hospital
• 著者所属(英): en
• 著者所属(英): Kyushu University Hospital
• 著者所属(英): en
• 著者所属(英): Kyushu Institute of Technology
• 著者名: Hoang, Anh Vy Ngo; Vu, Nguyen Phuong Quynh; Christina, Garcia; Shota, Fukushige; Hideki, Nakaguma; Naoki, Nakashima; Sozo, Inoue
• 著者名(英): Hoang, Anh Vy Ngo; Vu, Nguyen Phuong Quynh; Christina, Garcia; Shota, Fukushige; Hideki, Nakaguma; Naoki, Nakashima; Sozo, Inoue
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: In this paper, we recognize the activities of medical technicians using accelerometer data addressing the issues of timestamp duration and imbalanced data with segmentation technique and sampling. Research efforts mostly target patient and nurse activities while fewer studies are centered on medical technicians. Patient journey in hospitals often involves laboratory visits hence better comprehension of the medical technician routine is also vital. In this study, we investigate the activity recognition of 5 activities performed by a medical technician from Saiseikai Kumamoto Hospital by varying the overlapping segments and removing long-duration data. We compare the performance of balanced and imbalanced data. A 92% accuracy with 82% F1-score were achieved for the imbalanced data using 90 seconds window size and 40% overlapping using Random Forest algorithm while 84% accuracy with 78% F1-score for the balanced data using similar algorithm with 50% overlapping, removal of long-duration activities and random sampler.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: In this paper, we recognize the activities of medical technicians using accelerometer data addressing the issues of timestamp duration and imbalanced data with segmentation technique and sampling. Research efforts mostly target patient and nurse activities while fewer studies are centered on medical technicians. Patient journey in hospitals often involves laboratory visits hence better comprehension of the medical technician routine is also vital. In this study, we investigate the activity recognition of 5 activities performed by a medical technician from Saiseikai Kumamoto Hospital by varying the overlapping segments and removing long-duration data. We compare the performance of balanced and imbalanced data. A 92% accuracy with 82% F1-score were achieved for the imbalanced data using 90 seconds window size and 40% overlapping using Random Forest algorithm while 84% accuracy with 78% F1-score for the balanced data using similar algorithm with 50% overlapping, removal of long-duration activities and random sampler.
• 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: AA1271737X
• 書誌情報: 研究報告高齢社会デザイン（ASD）; 巻 2022-ASD-25, 号 2, p. 1-4, 発行日 2022-11-25
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2189-4450
• 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