From: Flexible human action recognition in depth video sequences using masked joint trajectories
(f) | (n) | (o) | (p) | (q) | (r) | (s) | (t) | |
---|---|---|---|---|---|---|---|---|
(f) | 81.82 % (9/11) | 18.18 % (2/11) | ||||||
(n) | 100 % (15/15) | |||||||
(o) | 90.91 % (10/11) | 9.09 % (1/11) | ||||||
(p) | 100 % (15/15) | |||||||
(q) | 100 % (15/15) | |||||||
(r) | 100 % (15/15) | |||||||
(s) | 100 % (15/15) | |||||||
(t) | 28.57 % (4/14) | 71.43 % (10/14) |