Abstract | ||
---|---|---|
Beautiful gait provides attractiveness and health, and some people want to know whether their gait is beautiful or not. Though methods of beautiful gait evaluation are proposed in previous works by utilizing 3D motion capture and force plates, they are not suitable for daily monitoring of beautiful gait. From this viewpoint, we propose a new method to estimate beautiful gait by utilizing machine learning classifiers with only one accelerometer on the body (chest, waist, or wrist). We conducted an experiment to verify the accuracy of our method with 41 subjects. Each subject wore accelerometers and walked on a straight path whose length was about 10 m 6 times. The accuracy of estimation in total 246 trials was calculated by two kinds of cross validation; leave-one-out cross validation and 41-fold cross validation. The former accuracy is 80.1% (chest), 82.1% (waist), and 82.9% (wrist). The latter accuracy is 73.2% (chest), 78.0% (waist), and 58.1% (wrist). Our contribution is to reveal the feasibility of estimation of beautiful gait using an accelerometer. |
Year | DOI | Venue |
---|---|---|
2016 | 10.1145/2968219.2971408 | UbiComp Adjunct |
Keywords | Field | DocType |
beautiful gait, accelerometer, wearable device | Motion capture,Computer vision,Wrist,Gait,Accelerometer,Waist,Computer science,Force platform,Artificial intelligence,Cross-validation,Gait evaluation | Conference |
Citations | PageRank | References |
0 | 0.34 | 5 |
Authors | ||
8 |
Name | Order | Citations | PageRank |
---|---|---|---|
Arinobu Niijima | 1 | 14 | 8.54 |
Kazuhiro Yoshida | 2 | 0 | 0.34 |
Osamu Mizuno | 3 | 51 | 12.30 |
Yumiko Tanmatsu | 4 | 0 | 0.34 |
naoki asanoma | 5 | 0 | 0.68 |
Tomoki Watanabe | 6 | 176 | 15.92 |
Tsubasa Nakayama | 7 | 0 | 0.34 |
Makoto Oyama | 8 | 0 | 0.34 |