Presentation 2018-03-01
A Method of Automatic Training Data Collection for Office Worker Seated-State using Machine Learning
Daisuke Ikeda, Yukihiro Tsuboshita, Takeshi Onishi,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Nowadays, approximately 75% of all employees in industrial countries are required to work in a seated position. According to a report, long-term seated position is suggested to have the relationship of health index and 28% of office workers have chronic low back pain. It is therefore important for office workers to take a break at an appropriate timing. To estimate office workers’ seated state automatically using Machine Learning, we must collect a lots of training data for obtaining a good performance. It requires however much labors and costs. In the present study, we present a method of automatic training data collection for Machine Learning using a smart phone attached to a person and accelerometers installed on a chair. On experiment of automatic training data collection, the average accuracy has shown 98.4% and this indicates that our method has reached a practical level.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Seated-state / Accelerometers / Machine Learning
Paper # LOIS2017-71
Date of Issue 2018-02-22 (LOIS)

Conference Information
Committee LOIS
Conference Date 2018/3/1(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Naha-City IT Souzoukan(Okinawa)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Hiroyuki Nishi(Sojo Univ.)
Vice Chair Tomohiro Yamada(NTT)
Secretary Tomohiro Yamada(Nagasaki Univ.)
Assistant Motoi Okamoto(Research Organization of Information and Systems)

Paper Information
Registration To Technical Committee on Life Intelligence and Office Information Systems
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Method of Automatic Training Data Collection for Office Worker Seated-State using Machine Learning
Sub Title (in English)
Keyword(1) Seated-state
Keyword(2) Accelerometers
Keyword(3) Machine Learning
1st Author's Name Daisuke Ikeda
1st Author's Affiliation Fuji Xerox Co., Ltd.(Fuji Xerox)
2nd Author's Name Yukihiro Tsuboshita
2nd Author's Affiliation Fuji Xerox Co., Ltd.(Fuji Xerox)
3rd Author's Name Takeshi Onishi
3rd Author's Affiliation Fuji Xerox Co., Ltd.(Fuji Xerox)
Date 2018-03-01
Paper # LOIS2017-71
Volume (vol) vol.117
Number (no) LOIS-464
Page pp.pp.1-5(LOIS),
#Pages 5
Date of Issue 2018-02-22 (LOIS)