Table 1.

The description of the symbols

Symbols	Descriptions
J	Jump length in input frames
$N_F$	Number of representative frames
$f_k$	$k^{\mathit{th}}$ representative frame
$v_i(u)$	Value of pixel u in block i
$N{G}_i(u)$	Spatial neighbourhood of pixel u in block i
$sk=\{f_1,f_2,...,f_{N_F}\}$	The sequence of skeleton with $N_F$ frames
$p^l=[p_1,...,p_I]$	Output feature vector of the deep network and input of classifications
V	Video activity
TF(.)	Conversion function