. 2019 Dec 6;9(12):1089. doi: 10.3390/ani9121089

Table 4.

Pearson correlation coefficients (r) between heart rate measurements of individual animals obtained from Polar monitors and, heart rate measured using computer vision techniques using different cameras (GoPro, Raspberry Pi), different areas on the animal (eye area, forehead, face) and with cameras located within different distances from the animal (1.5 m, 0.5 m). Data are the average for the animal within each period of handling (six handling periods per animal).

Invasive Method	Computer-Vision Method				Mean Correlation Coefficient (r) *	Range (r)	p-Value **
Invasive Method	Camera	Position	Distance	Analysed Area	Mean Correlation Coefficient (r) *	Range (r)	p-Value **
Polar monitor	RaspberryPi	Front	1.5 m	Eye area	0.89 ± 0.09	0.72–0.99	<0.05
Polar monitor	RaspberryPi	Front	1.5 m	Forehead	0.62 ± 0.23	0.32–0.90	<0.05
Polar monitor	RaspberryPi	Front	1.5 m	Face	0.16 ± 0.2	−0.11–0.39	>0.05
Polar monitor	RaspberryPi	Side	0.5 m	Eye area	0.78 ± 0.04	0.74–0.84	<0.01
Polar monitor	RaspberryPi	Side	0.5 m	Forehead	0.71 ± 0.18	0.53–0.89	<0.05
Polar monitor	GoPro	Side	0.5 m	Eye area	0.75 ± 0.14	0.55–0.92	<0.05
Polar monitor	GoPro	Front	0.5 m	Forehead	0.65 ± 0.08	0.58–0.76	<0.05

* Correlation coefficients are presented as mean ± SD; ** p-value are the highest observed in each category.