Researchers at the University of Washington have developed a groundbreaking solution called FeverPhone, which allows off-the-shelf smartphones to function as medically accurate body thermometers using only a new app. Device manufacturers have faced challenges in incorporating temperature sensors into smartphones and smartwatches for this purpose, but the FeverPhone app offers a promising solution.
While recent smart wearables like the Apple Watch Series 8 and Apple Watch Ultra now include sensors to measure body temperature, Apple maintains that the feature is not yet precise enough for medical diagnoses or treatments. Instead, these devices primarily utilize temperature measurements to provide users with insights into their sleep patterns. Despite offering heart rate readings, the Apple Watch does not serve as a digital thermometer capable of accurate temperature readings on demand.
During the COVID-19 pandemic, when demand for non-contact digital thermometers surged, availability became a significant issue. To address this, researchers at the University of Washington turned to smartphones as a readily accessible alternative. Unlike traditional digital thermometers, this solution does not require additional attachments or hardware upgrades. Smartphones already incorporate thermistors, which measure the device’s internal temperature, including the battery, to trigger safety precautions against overheating. This is why iPhones sometimes display warnings, advising users to allow the device to cool down before use.
Although thermistors, found in both smartphones and medical-grade thermometers, cannot directly measure a user’s body temperature while inside a phone, they can track the amount of heat energy transferred between the user and the mobile device during contact. To simulate a fevered test subject, the researchers heated a plastic bag full of water using a sous-vide machine. They then pressed the touchscreens of various smartphones against the heated bag, including devices in protective cases and those with screen protectors. By using the built-in thermistor to measure how rapidly the device warmed up during this interaction, the researchers collected data to train a machine-learning model that powers the FeverPhone app. This app can now estimate a user’s body temperature based on the gathered information.