AI-driven adaptive achromatic microscopy

In fluorescence microscopy, biological samples are often labeled with multiple fluorescent structures and thus require imaging at multiple wavelengths. This may lead to chromatic aberration, resulting from the shift of the focal spots of different wavelengths and causing worsened resolution. We present a correction of chromatic aberration by using an AI-driven adaptive achromatic lens. The adaptive achromatic lens consists of two lens chambers filled with different optical oils and is sealed with deformable glass membranes. By deforming the membranes, the focal spots can be re-adjusted to correct the chromatic aberration. An optical setup based on holography and a trained Reinforcement Learning agent is used to control the adaptive achromatic lens for investigation of biological samples. In experiments, it is proven that this approach can successfully correct system and sample induced aberrations and improve the imaging quality. Any commercial achromatic doublet does only correct the chromatic aberration induced by the system. In our research, the adaptive achromatic lens can reduce an initial chromatic aberration of system and sample of more than 0.07Dpt to below 0.0001Dpt.