Hand Posture Classification based on Surface Electromyography

hand-postures

This project classify hand posture using sEMG signals.

1. Segmentation

Number of channel Number of class Sampling frequency: 12
Total experiment time (All classes): 17
Repetition (Session): 2,000 Hz
Time per repetition: 15 min (1,800,000 points) 6 sessions per class 3-5s
Duration per repetition : 3~5 s
Segment size : 200 ms

2. 통계적 수치 계산 (Time domain feature)

1개의 신호 구간에서 36개의 통계적 수치를 계산한다. → [MAV, VAR, WL] x 12 (channels)

$$MAV = \frac{1}{N}\sum_{i=1}^{N} |x_{i}|$$

$$VAR = \frac{1}{N}\sum_{i=1}^{N} |x_{i}^{2}|$$

$$WL = \sum_{i=1}^{N} |x_{i+1} - x_{i}|$$

3. Train / Test

Repetition 1 ~ 4: train set
Repetition 5 ~ 6: test set

4. Result

Accuracy: 55%