Miniaturization and expansion of the contactless temperature measurement system. Facial temperatures in relation to age, pulse and gender.

The dataset contains temperature measurements on the surface of the face taken on 109 people. Each patient (identified by Patient ID in dataset) acclimatized in a room with a temperature of 22-24 degrees Celsius. Then the person completed a survey, during which they provided their:

• age (column Survey - age [years]),
• gender (column Survey - Gender),
• temperature measurement using a pyrometer thermometer (column Survey - temperature [°C]),
• and pulse measurement using a pulse oximeter (column Survey - measured pulse [BPM]).

After that, the examined person stood in front of the contactless temperature measurement system (using a thermal camera), which was continuously calibrated to the black body at a distance of 1.5-3 meters (column Distance between camera and patient [m]). Then, several hundred temperature measurements were taken on each person in the following ways:

• Median temperature on face [°C]
• Median temperature on face, 1% of pixels with max temperature [°C]
• Median temperature on face, 5% of pixels with max temperature [°C]
• Median temperature on face, 10% of pixels with max temperature [°C]
• Median temperature in the center of the eyes (3x3 pixels) [°C]
• Median temperature measured at the corners of the eyes (3x3 pixels) [°C]

Additionally, the system automatically estimated:

• the age of the examined person (column Estimated Age [years]),
• the pulse of the examined person (column Estimated Pulse [BPM]),
• and gender (Estimated Gender).

According to [1], the measured temperature on the surface of the face is influenced by the age of the measured person. As part of the project, a Binary Regression Tree was developed, which considers (estimated) age when calculating the temperature on the surface of the face (column Temperature calculated by Binary Tree Regression algorithm [°C]).

[1] Cheung, Ming & Chan, Lung & Lauder, I & Kumana, Cyrus. (2012). Detection of body temperature with infrared thermography: accuracy in detection of fever. Hong Kong medical journal = Xianggang yi xue za zhi / Hong Kong Academy of Medicine. 18 Suppl 3. 31-4.