| Project/Area Number |
18H04115
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (A)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Review Section |
Medium-sized Section 61:Human informatics and related fields
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
Yagi Yasushi 大阪大学, 産業科学研究所, 教授 (60231643)
|
|---|
| Project Period (FY) |
2018-04-01 – 2021-03-31
|
| Project Status |
Completed (Fiscal Year 2020)
|
| Budget Amount *help |
¥44,460,000 (Direct Cost: ¥34,200,000、Indirect Cost: ¥10,260,000)
Fiscal Year 2020: ¥13,910,000 (Direct Cost: ¥10,700,000、Indirect Cost: ¥3,210,000)
Fiscal Year 2019: ¥12,740,000 (Direct Cost: ¥9,800,000、Indirect Cost: ¥2,940,000)
Fiscal Year 2018: ¥17,810,000 (Direct Cost: ¥13,700,000、Indirect Cost: ¥4,110,000)
|
|---|
| Keywords | コンピュータビジョン / パターン認識 / バイオメトリクス |
|---|
| Outline of Final Research Achievements |
We developed a comprehensive gait model including four factors: skill (or experience, capability), emotion (or feeling), intention (or situation), body (or age, health condition, disease, impaired body function), under a hypothesis where any gaits can be represented by a composite of the four factors We constructed several gait databases including the factors, e.g., a gait dataset with a health indicator, i.e., body composition as a body label. We then design a deep neural network which is pre-trained and fine-tuned across multiple factors and showed its effectiveness in a task of body composition estimation from a gait video. We also introduced insight obtained through the comprehensive gait model construction into surveillance applications, and developed a gait recognition framework which employs disentangled representation learning of individuality and other covariate factors similarly to we disentangled a gait into the above-mentioned four factors.
|
| Academic Significance and Societal Importance of the Research Achievements |
本研究は、歩行解析の新しい枠組みの 提案であり、情報学的観点から歩行を科学する研究で学術的価値が大きい。
映像からの歩行解析では、個人認証の応用としての科学捜査支援が主流であったが、知情意体 のモデル化により、歩行映像から、個人認証だけでなく、感情、意図、健康度が同時推定できる。 このような技術は、日常歩行をカメラにより見守るだけで、異常の早期発見が可能となり、家庭、 学校、職場など人々が日常暮らす様々な場面で、防犯とメンタルケア、ヘルスケア、さらに、個 別サービスを同時に実現できる
|
