Indirect sensing |
Video cameras (33, 36, 63) |
1. Easy to understand |
1. Computationally expensive analysis |
|
|
2. High spatial resolution |
2. Privacy concerns |
|
|
3. High context information |
3. Large disk space requirements |
|
|
4. Portable |
4. Generally low temporal resolution |
|
|
5. High availability |
5. Occlusion issues |
|
3D motion capture (29, 37) |
1. High spatial resolution |
1. High costs |
|
|
2. Depth information |
2. Computationally very expensive analysis |
|
|
3. Accurate motion capture |
3. Privacy concerns |
|
|
4. High reliability |
4. Very large disk space requirements |
|
|
5. High temporal resolution possible |
5. Large physical space requirement |
|
|
6. Secondary movement analysis possible (such as force and weight exchange) |
6. Markers needed for motion capture |
|
|
|
7. Occlusion issues |
|
Microsoft kinect (35, 47, 62) |
1. High spatial resolution |
1. Not suitable for infants (<4 years) |
|
|
2. Depth information |
2. Occlusion issues |
|
|
3. Low-cost |
3. Low temporal resolution |
|
|
4. Marker-less motion capture |
4. Limited field of view |
Direct sensing |
Wearable movement sensors (11–13, 34, 39–43, 45, 64) |
1. High temporal resolution |
1. Low spatial resolution |
|
|
2. Low-cost |
2. Occasional data losses (wireless) |
|
|
3. Energy efficient |
3. Limited battery life (wireless/real-time) |
|
|
4. Privacy preserving |
4. Difficulty in consistent positioning |
|
|
5. Small physical size |
5. Comfort issues |
|
|
6. Good battery life (embedded) |
6. Relative movement capture only |
|
|
7. High availability (e.g., mobile phones) |
|
|
|
8. Actigraphs: sleep/wake patterns |
|
|
Magnet tracking system (31, 46, 66) |
1. High temporal resolution |
1. High costs compared with accelerometers |
|
|
2. Very high accuracy |
2. Computationally very expensive analysis |
|
|
3. Metal tolerant |
3. Complex setup |
|
|
4. No line of sight occlusions |
4. Magnetic and electrical interference issues |