| Project/Area Number |
19K24364
|
|---|
| Research Category |
Grant-in-Aid for Research Activity Start-up
|
| Allocation Type | Multi-year Fund |
|---|
| Review Section |
1002:Human informatics, applied informatics and related fields
|
|---|
| Research Institution | Osaka University |
|---|
Principal Investigator |
ALLAM ALLAM 大阪大学, 産業科学研究所, 特任研究員(常勤) (70850767)
|
|---|
| Project Period (FY) |
2019-08-30 – 2021-03-31
|
| Project Status |
Completed (Fiscal Year 2020)
|
| Budget Amount *help |
¥520,000 (Direct Cost: ¥400,000、Indirect Cost: ¥120,000)
Fiscal Year 2020: ¥260,000 (Direct Cost: ¥200,000、Indirect Cost: ¥60,000)
Fiscal Year 2019: ¥260,000 (Direct Cost: ¥200,000、Indirect Cost: ¥60,000)
|
|---|
| Keywords | Gait recognition / Multi objects tracking / Optical flow / Dense Trajectory / FIsher vecotr encoding / Multi-Gait Modeling / Gait Energy Image / Feature Representation / Dense Trajectories / Fisher vector encoding |
|---|
| Outline of Research at the Start |
This work introduces a novel proposition toward a multi-gait analysis/recognition system. Instead of existing single gait-based systems, I propose to recognize the identities of multi walking persons in the same videos.As well, I will study the impacts of multi walking persons on their gaits styles.
|
| Outline of Final Research Achievements |
Through our this research project, I finished the following tasks:
1- Creating the single gait dataset which containing 182 walking subjects. 2- Compiling the multi-gait dataset from my lab recordings of group walking experiment. 3- Extract the optical flow and dense trajectories of both the single and multi gait dataset. 4- I used the optical flow and dense trajectory to build local motion descriptors for each walking subject by considering the aggregation of the optical flow motion information along with each extracted dense trajectory. As well, I computed the relative position and shape descriptor for each trajectory. 5- To build a global motion descriptor for each subject given the calculated local descriptor, I applied the robust fisher vector encoding technique. To evaluate the proposed feature representation, I measured the pairwise similarity between single and multi global descriptor for each subject. I used it as well in gait relative experiment.
|
| Academic Significance and Societal Importance of the Research Achievements |
I introduced the first steps toward a multi-gait recognition system. Ihave prepared the first multi-gait dataset for research purposes. The dataset contains, videos, optical flow, BBXs,and dense trajectories for all subjects.I introduced a robust motion feature representation for model evaluation.
|
