UCF-IPhone Data Set
| Paper: | Macro-Class Selection for Hierarchical K-NN Classification of Inertial Sensor Data |
| Data Set: | Click here |
| Presentation: | Click here |
Publication
Corey McCall, Kishore Reddy and Mubarak Shah, Macro-Class Selection for Hierarchical K-NN Classification of Inertial Sensor Data, Second International Conference on Pervasive and Embedded Computing and Communication Systems, PECCS 2012, February 24-26, 2012, Rome, Italy.
Overview
Aerobic actions were recorded from subjects using the Inertial Measurement Unit (IMU) on an Apple iPhone 4 smartphone. The IMU includes a 3D accelerometer, gyroscope, and magnetometer*. Each sample was taken at 60Hz, and manually trimmed to 500 samples (8.33s) to eliminate starting and stopping movements. iPhone is always clipped to the belt on the right hand side as shown in the picture.
Data Set Details
| Actions | Number of Actors | Total Number of Instances |
| Biking | 6 | 30 |
| Climbing Stairs | 9 | 45 |
| Descending Stairs | 9 | 45 |
| Gym Biking | 8 | 39 |
| Jump Roping | 9 | 45 |
| Running | 9 | 45 |
| Standing | 9 | 45 |
| Treadmill Walking | 9 | 44 |
| Walking | 9 | 45 |
Statistics
Note: Our experiments show that the data from magnetometer is not useful. After feature selection was performed, the most useful features came dominantly from the accelerometer data.