Facial Pain Expression Recognition in Real-Time Videos

Table 1

Performance analysis of 22 videos.

Video number	Video size (MB)	Video format	Number of frames	Training time	Testing time	HR (pain) in range (%)
1	3.02	.mp4	140	7 m 5 s	2 m	80.89–86.41
2 (live)	72	.avi	232		2 m	90.47–92.46
3 (live)	116.5	.avi	512		5 m	74.02–82.23
4	102.2	.wmv	1011	15 m	4 m	91.12–99.34
5	7.95	.mp4	200	4 m 56 s	1 m	90.00–98.76
6	100.2	.mov	700	9 m	3 m 36 s	69.11–74.29
7	167.4	.avi	1072	16 m	7 m	82.92–87.85
8 (live)	234	.avi	1598		8 m 53 s	70.17–78.76
9	9.47	.mp4	223	5 m 18 s	1 m 53 s	87.77–90.00
10	700	.avi	2236	20 m 19 s	5 m	85.55–91.72
11	7.9	.mov	70	20 s	5 s	91.00–93.33
12	28.8	.mov	220	45 s	11 s	78.76–82.92
13	50	.avi	250	9 m 18 s	2 m 40 s	94.33–100.00
14	30.2	.mov	231	3 m 39 s	1 m 11 s	92.09–97.56
15 (live)	136.1	.avi	754		5 m 30 s	65.88–69.11
16	42.6	.wmv	470	7 m	2 m 43 s	79.14–81.21
17	46.1	.wmv	501	8 m 13 s	3 m	83.44–87.11
18	50.4	.wmv	540	9 m 27 s	3 m 11 s	75.34–86.21
19	10.1	.mp4	219	11 m 7 s	1 m	67.45–75.46
20	30.2	.avi	209	8 m 43 s	1 m 34 s	67.11–70.30
21	352.2	.avi	913	15 m 19 s	5 m	79.11–84.29
22 (live)	194.5	.avi	1041		6 m 17 s	69.90–74.87