Globally deployed COVID-19 fever screening devices using infrared thermographs consistently normalize high readings to afebrile range

Conor Healy; Zachary Segal; Chris Hinnerichs; Ethan Ace; Derek Ward; John Honovich

doi:10.1117/1.JBO.26.4.043009

Translator Disclaimer

13 March 2021 Globally deployed COVID-19 fever screening devices using infrared thermographs consistently normalize high readings to afebrile range

Conor Healy, Zachary Segal, Chris Hinnerichs, Ethan Ace, Derek Ward, John Honovich

Author Affiliations +

J. of Biomedical Optics, 26(4), 043009 (2021). https://doi.org/10.1117/1.JBO.26.4.043009

Table 1

Devices stated accuracy and emissivity by type.

IRTs, NCIT, and blackbodies used	Type	Stated accuracy	Stated emissivity
Extech IR200: noncontact forehead infrared thermometer	NCIT	$\pm 0.3 ° C$	NA
FLIR E54-EST thermal screening solutions	IRT	$\pm 0.3 ° C$	NA
TVT TD-E2128-TM face recognition terminal	IRT	$\pm 0.3 ° C$	NA
Bems temperature terminal	IRT	Not stated	NA
Dahua DH-TPC-BF5421-T thermal hybrid network camera	IRT	$\pm 0.33 ° C$	NA
ZKTeco SF1008+ body temperature + mask detection access control reader	IRT	$\pm 0.33 ° C$	NA
Certify SnapXT Pro	IRT	$\pm 0.3 ° C$	NA
Hikvision DS-2TD2636B-13/P temperature screening thermographic camera	IRT	$\pm 0.5 ° C$	NA
Meridian clear 2 temperature screening Kiosk	IRT	$\pm 0.5 ° C$	NA
Dahua JQ-D70Z blackbody	BB	$\pm 0.2 ° C$	0.97
Sunell SN-TH01 portable blackbody	BB	$\pm 0.2 ° C$	0.97

Fig. 1

View Large | Download | View In Article Context

Testing configuration for simulated human thermographic reading. (a) Regular image. (b). Thermal image. Subject was mounted at 168 cm from the floor. Manufacturer specifications were followed in the placement of the IRTs relative to the subject. All tablets were mounted at a height of 155 cm and placed at a distance of 50 cm from the subject. The FLIR IRT was mounted at a height of 157.5 cm and placed 70 cm from the subject.

Fig. 2

View Large | Download | View In Article Context

IRT image of (a) male with normal temperature and (b) male with elevated temperature from IEC 80601.3

Fig. 3

View Large | Download | View In Article Context

Testing environment. (a) Side view. (b) Top view.

Fig. 4

View Large | Download | View In Article Context

Testing environment for devices requiring BB calibration with subject and reference device placed side-by-side. (a) Side view. (b) Top view.

Table 2

Deviation from control temperature reading.

Device	Range (°C)	Variance °C	Mean °C	Device versus subject temp.		Difference between device and subject versus subject temp.
Device	Range (°C)	Variance °C	Mean °C	Adjusted R-squared	Regression coefficient	Adjusted R-squared	Regression coefficient
Control devices
FLIR IRT	35 to 40	2.4	37.5	0.99	1	0	−0.001
Extech surface NCIT	35.05 to 40.05	2.4	37.5	0.99	1	0	−0.001
Extech body NCIT	36.7 to 40.2	1.3	38.2	0.98	0.78	0.8	−0.22
Tested devices
Hikvision Bullet IRTa	35.6 to 39.6	1.6	37.4	0.99	0.8	0.88	−0.19
Dahua Bullet IRTb	35.4 to 39.4	1.4	37.3	0.99	0.77	0.95	−0.23
Meridian IRTc	36.9 to 39.8	0.95	37.8	0.91	0.6	0.82 (.91 log)d	−0.4
ZKTeco IRT	36.3 to 39.0	0.85	37.3	0.94	0.58	0.89	−0.42
Certify IRT	36.4 to 39.0	0.72	37.3	0.93	0.53	0.91	−0.47
TVT IRT	36.6 to 37.8	0.16	36.9	0.7	0.22	0.97	−0.78
Bems IRT	36.5 to 36.8	0.005	36.7	0.001	0	0.99	−1

Fig. 5

View Large | Download | View In Article Context

Bland–Altman plot of seven tested devices.

Fig. 6

View Large | Download | View In Article Context

Recorded readings for tested IRT reading versus subject.

Fig. 7

View Large | Download | View In Article Context

Bland–Altman plot of Hikvision DS-2TD2636B-13/P Results.

Fig. 8

View Large | Download | View In Article Context

Bland–Altman plot of Dahua DH-TPC-BF5421-T.

Fig. 9

View Large | Download | View In Article Context

Bland–Altman plot of Meridian.

Fig. 10

View Large | Download | View In Article Context

Bland–Altman plot of ZKTeco.

Fig. 11

View Large | Download | View In Article Context

Bland–Altman plot of Certify.

Fig. 12

View Large | Download | View In Article Context

Bland–Altman plot of TVT.

Fig. 13

View Large | Download | View In Article Context

Bland–Altman plot of Bems.

Fig. 14

View Large | Download | View In Article Context

Disparity between control (FLIR) and tested IRTs: magnitude and direction of deviation from BB reading. (a) Difference between device and subject versus subject. (b) Absolute value of difference between device and subject versus subject.

Fig. 15

View Large | Download | View In Article Context

Pooled differences between tested devices and BB temperature.

Fig. 16

View Large | Download | View In Article Context

Results from control devices, FLIR E54-EST and Extech IR200.

Table 3

Variance across measured temperatures by control and tested devices.

Device	Variance (°C)	Mean (°C)	Range (°C)
Hypothetical data
BB	2.4	37.5	35 to 40
Perfect device with no offset	2.4	37.5	35 to 40
Perfect device with a fixed offset of +1	2.4	38.5	36 to 41
An inaccurate device that does not normalizea	3.3	37.5	34.4 to 40.7
Experimental data
FLIR	2.4	37.5	35 to 40
Extech surface mode	2.4	37.5	35.05 to 40.05
Extech body mode	1.3	38.2	36.7 to 40.2
Hikvisionb	1.6	37.4	35.6 to 39.6
Dahuac	1.4	37.3	35.4 to 39.4
Meridiand	0.95	37.8	36.9 to 39.8
ZKTeco	0.85	37.3	36.3 to 39.0
Certify	0.72	37.3	36.4 to 39.0
TVT	0.16	36.9	36.6 to 37.8
Bems	0.005	36.7	36.5 to 36.8