| Budget Amount *help |
¥4,810,000 (Direct Cost: ¥3,700,000、Indirect Cost: ¥1,110,000)
Fiscal Year 2017: ¥1,560,000 (Direct Cost: ¥1,200,000、Indirect Cost: ¥360,000)
Fiscal Year 2016: ¥650,000 (Direct Cost: ¥500,000、Indirect Cost: ¥150,000)
Fiscal Year 2015: ¥2,600,000 (Direct Cost: ¥2,000,000、Indirect Cost: ¥600,000)
|
|---|
| Outline of Final Research Achievements |
This research aims at developing activity recognition techniques using wearable sensors for various applications. It focuses on the feature extraction phase. It has experimentally evaluated how the performance of activity recognition from acceleration data depends on the ways of extracting features using denoising autoencoders. The experimental results show the effects on classification accuracy of the combinations of the body parts from which the acceleration data are obtained, the ways of providing a stacked denoising autoencoder with the acceleration data, the different numbers of nodes in the output layer of each stacked denoising autoencoders, and the different sizes of the time windows from which the acceleration data are extracted.
|
