Data-driven super-resolution optoacoustic imaging via physically encoded signal acquisition

Super-resolution optoacoustic (OA) imaging, enabled by localization and tracking of highly absorbing circulating microparticles, has overcome the longstanding resolution–depth trade-off limiting optical-contrast imaging methods. Responses recorded from individual microparticles further allow for calibrating the spatially dependent OA impulse response, which can support accurate data-driven learning of a linear forward model for tomographic reconstructions, even in complex, physically encoded acquisition schemes. We demonstrate accurate localization of microparticles by exploiting multiple ultrasound scattering to physically encode absorber positions in time-resolved OA signals. Using a model calibrated via raster scanning of a microsphere, we achieved in vivo localization and tracking of intravenously injected microparticles, enabling localization optoacoustic tomography (LOT) with over an order-of-magnitude fewer transducer elements than previously employed. Furthermore, auxiliary-transducer-assisted localization of microspheres in the bloodstream facilitated implementation of a self-calibration methodology, allowing for super-resolution imaging with a single time-resolved signal.