Abstract:Skin exposome encapsulates all internal and environmental exposures that affect skin health. Of these, photo-pollution refers to the combined effect on human skin of the simultaneous exposure to solar radiation (especially UV) and air pollution. Providing personalised photo-pollution exposure warnings and dose monitoring to an individual through a smartphone app could help in reducing skin ageing and degradation as well as in managing skin conditions (for example Atopic Dermatitis). However, accurate monitoring is challenging without a potentially expensive or cumbersome sensor device. In this work we present an innovative satellite-based air pollutant monitoring software service, ExpoPol, developed by siHealth Ltd. ExpoPol synthesises several inputs including live satellite imagery in real-time into an artificial intelligence (AI) model to provide assessment of the exposure of a smartphone user to relevant air pollutants, such as nitrogen oxides (NO x ), poly-aromatic hydrocarbons (PAH) and ozone (O 3 ). When combined with siHealth's patented technology HappySun® for solar radiation monitoring, ExpoPol can effectively provide a sensor-less personal skin photo-pollution dosimetry. By downscaling satellite data using local geographic data, ExpoPol is capable of monitoring pollutants with street-level resolution and global coverage in near real-time. We evaluate the accuracy of ExpoPol against ground-station monitoring data for three pollutants across three continental regions (Europe, Asia, North America) and find R 2 values of 0.62, 0.65, 0.74 for PM 10 , PM 2.5 , NO 2 respectively. ExpoPol is shown to be significantly more accurate than a state-of-the-art global atmospheric forecasting system (CAMS) over the same ground-station dataset and provide data on much finer spatial resolutions. The presented system can support the real-time automatic assessment of the user's skin exposome, anywhere and anytime. This paves the way for the development of mobile applications empowering users and clinicians to make informed decisions about skin health, or assisting dermocosmetic manufacturers in the creation of personalised products for personal care (e.g., skin ageing prevention or hair care).

A satellite-based, near real-time, street-level resolution air pollutants monitoring system using machine learning for personalised skin health applications