Exploiting Serialized Fine-Grained Action Recognition Using WiFi Sensing

Table 1

Time (T) and frequent (F) domain features used for stroke recognition.

ID	Features	IG
1	Mean absolute deviation (T)	2.161
2	RMS amplitude (T)	2.110
3	Mean (T)	2.071
4	Median (T)	2.060
5	Maximum (T)	1.856
6	Minimum (T)	1.839
7	Standard deviation frequency (F)	1.782
8	Mean frequency (F)	1.086
9	Kurtosis coefficient (T)	0.825
10	Magnitude margin factor (T)	0.783
11	Peak factor (T)	0.737
12	Impulse factor (T)	0.731
13	Variance (T)	0.649
14	Standard deviation (T)	0.643
15	Form factor (T)	0.628
16	Peak-to-peak value (T)	0.585
17	Kurtosis factor (T)	0.534
18	RMS frequency (F)	0.507
19	Skew coefficient (T)	0.446
20	Skewness factor (T)	0.403
21	Variance frequency (F)	0.403
22	Energy (F)	0.393
23	Central frequency (F)	0.391
24	Quartile deviation (T)	0.385