MEMS Devices-Based Hand Gesture Recognition via Wearable Computing

. 2023 Apr 27;14(5):947. doi: 10.3390/mi14050947

Algorithm 1: Assisted Segmentation Algorithm

Input: Continuous gesture data: $D = \{d_{1}, d_{2}, d_{3}, . . ., d_{n}\}$ ;

Output: Split start point list S, split end point list O.

1
Transition state mean: $m = \frac{\sum_{i = 1}^{100} d_{i}}{100}$ ;
2
Define the length l of the sliding window sequence, the current window value $A_{0}$ , the window value $A_{1}$ after sliding 5 times, the window value $A_{2}$ after sliding 10 times, $E_{1} = |A_{0} - A_{1}|$ , $E_{2} = |A_{1} - A_{2}|$ ;
3
Traverse continuous gesture data D;
4
if $(E_{1} > 10) & & (|A_{0} - m| < 12) & & (|A_{2} - m| < 12)$ then Read the current position and add it to the starting point list S;
5
if $(E_{2} > 10) & & (|A_{0} - m| < 12) & & (|A_{2} - m| < 12)$ then Read the current position and add it to the end point list O;
6
End;