Label-free bio-aerosol detection and classification using a virtual impactor, holography and deep learning

We present a computational mobile imaging device that captures in-line holograms of bioaerosols flying through a virtual impactor. Using deep neural networks, this cost-effective system can perform long-term air quality monitoring, automatically identifying/counting various allergens.