Language：Japanese   Publisher：Faculty of Design, Kyushu University  

In behavioral observation within videos using keypoint detection technology, the movement of specific points (keypoints) is tracked to quantify behavior. The use of video cameras enables long-term observation and simultaneous observation of multiple people, but this increases the burden on observers. Therefore, the efficiency and automation of behavioral recording using video analysis technology have become important challenges for the future. The author is conducting research to expand quantitative methods of behavioral observation by combining video cameras with keypoint detection technology. There are two main challenges being addressed in this research. The first is improving the accessibility of keypoint detection technology. Despite technological advancements, the latest technologies are difficult to use for non-information technology students and researchers. To address this, the author is developing an intuitive application designed to handle various image recognition libraries. The second challenge is converting keypoints into meaningful behavioral data. An application is being developed to transform keypoint coordinate data into behavioral data, allowing for analysis of movement speed and direction of keypoints. Currently, these technologies are being implemented into applications to verify their practical applicability in behavioral observation.

DOI： 10.15017/7170833

CiNii Research

Language：Japanese  

Questionnaires and behavioral observation have long been used in design methods. However, knowledge of human behavior has not been adequately applied to design methods from either a quantitative or qualitative perspective. Therefore, we are aiming to develop a resilient design method for extracting and solving physical, physiological, and psychological challenges faced by users, such as operational issues associated with images recorded by security cameras installed at disaster evacuation centers. The extracted issues and their solutions must then be incorporated into the operational design. We conducted an experiment simulating activity at an evacuation center and distributed a questionnaire survey to participants. We analyzed HRV when participants were grouped together in spaces according to the F-formation system in order to determine what constitutes stressful conditions. The results of this pilot study revealed a relationship of grouping arrangement, personality traits, and HRV among the participants. It was demonstrated that the experimental results could be applied to the resilient design of disaster evacuation centers.

DOI： 10.15017/2349290

Language：Japanese  

本論文では，セラミックス製品の表面に形成する nearly invisible な 2 次元コードの形成方法と，当該コードから文字情報（公開情報と暗号化された秘密情報）を抽出する，スマートフォンと紫外線光源を用いた2 次元コードリーダシステムを紹介する．Nearly invisible な 2 次元コードは，セラミックス製品の意匠を可能な限り損なわずに文字情報を製品に持たせるために著者らが開発したものであり，当該コードは紫外線が照射されている間，青白い光を発光するため肉眼での視認が容易となる．リーダシステムは，カメラ機能と 2 次元コード認識アルゴリズムを用いて文字情報を抽出するが，デジタル鍵にアクセスできるスマートフォンのみが暗号化された秘密情報を復号できる．実験の結果，nearly invisible な 2 次元コードが形成できること，およびリーダシステムが文字情報を再現性をもって抽出できることがわかった．

Language：Japanese  

本展示では，意匠が重視される製品の 1 つとしてセラミックス製品に注目し，この表面に公開情報と暗号化された秘密情報を格納できるセキュアな 2 次元コード（SQRC）をnearly invisibleにして焼きつける方法と，焼き付けられた SQRC を明瞭化しつつ文字情報（公開情報と秘密情報）を抽出できるスマートフォンベースのリーダシステムを紹介する．製品に関する情報を公開情報，製品の真正性を検証する方法を秘密情報とすることで，製造業者のみが製品ごとの秘密情報をスマートフォンを使って読むことを可能にする．一方消費者は，製造者が公開している製品に関する情報に簡単にアクセスできるため，製造業者と製品に対する信頼と安心を深めることができる．nearly invisible な 2 次元コードが形成できること，およびリーダシステムが文字情報を再現性をもって抽出できることを，展示において紹介する．

Language：Japanese  

type:資料
Most of support to disaster victims are material support. It is very necessary to provide material support when a disaster happens. The disaster victims often have to surmount emerging problems and uncertainties in their new lives meanwhile making a living. Their self-equipped resilience of surmounting difficulties translates into the problem solving capacity. Generalization and summaries of the above are the basis for behavior modification and the possibility to establish a new social foundation. This kind of resilience would translate into effective action when disaster falls next time, which would greatly prevent or mitigate disasters. The resilience-oriented behavior modification would bring new changes to the society, and generate the power to create a new future. In light of design theory (industrial design, functional engineering, human engineering), this study aims to develop a methodology for resilience design.

DOI： 10.15017/1905828

Responsible for pages：総ページ数：xvii, 705p   Language：Japanese  

Event date： 2024.3

Language：Japanese   Presentation type：Oral presentation (general)  

Venue：九州大学大橋キャンパス   Country：Japan  

This study explores the effectiveness and challenges of quantitative behavioral observation using keypoint detection technology, specifically during smartphone use on beanbag chairs. We conducted experiments with 20 participants, divided into 10 pairs, in a controlled laboratory setting. Utilizing deep learning models for body keypoint detection, we analyzed the participants' postures and movements captured on video. Our findings indicate the potential of this technology for efficient, quantitative observation in product development, particularly in ergonomics and user interaction analysis. This research contributes to the

Language：Others  

Country：Other  

Language：Others  

Country：Other  

Language：Others  

Country：Other  

Language：Others  

Country：Other  

Language：Japanese  

DOI： 10.15017/7170833

Work type：Software  

Behavior Senpai is an application for conducting quantitative behavioral observations. It converts human behavior captured by a video camera into time-series coordinate data using keypoint detection AI, allowing for the quantitative analysis of human behavior using this time-series coordinate data.

DOI： 10.48708/7160651

researchmap

Other Link： https://hdl.handle.net/2324/7160651

Work type：Software  

researchmap

Work type：Software  

researchmap

Type：Competition, symposium, etc. 

Type：Peer review 

Number of peer-reviewed articles in foreign language journals：1