Skip to main content
. 2021 May 25;11(6):1549. doi: 10.3390/ani11061549

Table 5.

Summary of labeled crowd and eat/drink datasets.

No. Valid Videos Eat/Drink Dataset Crowd Dataset
#
Videos
#
Labels
Mean
Length (s)
Total
Length (H:M:S)
#
Videos
#
Labels
Mean
Length (s)
Total
Length (H:M:S)
5367 10,529 56 163:51:46 262 409 188 21:26:15
Behaviors
DRINK 1072 1586 12.6 5:33:14 151 311 5 0:25:52
EAT 1484 3063 43.4 36:57:31 260 375 65 6:46:33
LICKOBJECT 897 2056 5.6 3:11:45 121 212 5.7 0:20:18
LICKSELF 464 1405 12.4 4:50:11 8 8 7.3 0:00:58
SCRATCH 416 679 6.8 1:16:41 4 5 4.0 0:00:20
PETTING 510 901 6.7 1:41:03 72 92 2.0 0:04:04
RUBBING 271 605 7.0 1:10:32 0 0 0 0:00:00
SHAKE 552 689 1.7 0:19:09 48 64 1.0 0:1:04
SNIFF 2668 9696 3.9 10:31:56 241 1082 5.0 1:36:46
Postures
LIE DOWN 1340 2407 65.7 43:56:24 10 11 58.8 0:10:46
MIXED 4072 11,538 9.9 31:49:27 261 1492 30.0 12:44:22
VIGOROUS 633 1968 8.7 4:45:24 15 24 6.9 0:02:44
SIT 1500 2784 19.8 15:18:20 136 263 21.0 1:32:42
STAND 3725 9090 22.5 56:44:37 234 1294 19.0 6:55:51
WALK 1255 4101 9.9 11:16:34 5 21 6.0 0:02:06