Adaptive Optics Microscopy Using Physical Models for Aberration Correction

The patent applies preliminary action by using a wavefront sensor to measure and characterize the scattering body's distortion properties before the actual imaging process. This pre-characterization creates a distortion map that is stored and later used to correct images, eliminating the need for large training data sets. The wavefront sensor performs the measurement in advance, allowing the neural network to be trained on a small data set using these pre-measured distortion parameters.

The patent introduces a wavefront sensor as an intermediary device that measures the distortion caused by the scattering body. This intermediary provides direct physical measurements of the aberrations, which then serve as training data for the neural network. Instead of requiring the network to learn distortion patterns from large image data sets, the wavefront sensor acts as a mediator that provides precise, targeted measurements that dramatically reduce the training data requirements.

The patent applies universality by designing the wavefront sensor to serve multiple functions: it characterizes the scattering body's distortion, provides training data for the neural network, and enables real-time correction. The same optical path and components are used for both wavefront sensing and image acquisition, allowing the system to perform multiple tasks without requiring separate dedicated hardware for each function, thereby limiting the increase in complexity.

The system applies self-service by using the scattering body itself as the object of measurement for the wavefront sensor. The wavefront sensor measures the distortion introduced by the scattering body directly, and this measured distortion is then used to correct images of the same scattering body. The system essentially measures and corrects itself, eliminating the need for separate calibration objects or additional complexity in the measurement setup.

The patent applies merging by combining the measurement of excitation path aberrations and detection path aberrations into a single wavefront sensing operation. The wavefront sensor captures the total distortion affecting the optical path, and the neural network processes this combined information to simultaneously correct both excitation and detection path aberrations. This merging approach maintains measurement precision while dramatically reducing the time and complexity compared to separate measurements.

The system applies continuity of useful action by implementing real-time wavefront sensing and correction during the imaging process. Rather than performing separate calibration measurements before imaging, the wavefront sensor continuously monitors and characterizes the distortion during actual sample imaging, allowing the neural network to provide continuous correction. This eliminates idle measurement time and ensures the correction is always current and accurate.