View Code? Open in Web Editor
NEW
Human Activity Recognition (HAR). Computer Vision and Deep Learning.
human-activity-recognition's Introduction
Human Activity Recognition
- MPII Human Pose Dataset.
- UCI Human Activity Recognition Using Smartphones Data Set.
- Static: Pictures (Deep learning methodology: Convolutional Neural Networks).
- Dynamic: Smartphone sensor data (Machine learning methodology: Logistic Regression, Support Vector Machine, Random Forest).
- Use more epochs for training the last model.
- Increase the number of filters in the convolutional layers.
- Expand the dataset.
- For the dynamic approach, use different strategies and another dataset.
human-activity-recognition's People
Contributors
Watchers