Abstract:The application of machine learning in wavefront reconstruction has brought great benefits to real-time, non-invasive, deep tissue imaging in biomedical research. However, due to the diversity and heterogeneity of biological tissues, it is difficult to train the dataset with a unified model. In general, the utilization of some unified models will result in the specific sample falling outside the training set, leading to low accuracy of the machine learning model in some real applications. This paper proposes a sensorless wavefront reconstruction method based on transfer learning to overcome the domain shift introduced by the difference between the training set and the target test set. We build a weights-sharing two-stream convolutional neural network (CNN) framework for the prediction of Zernike coefficient, in which a large number of labeled randomly generated samples serve as the source-domain data and the unlabeled specific samples serve as the target-domain data at the same time. By training on massive labeled simulated data with domain adaptation to unlabeled target-domain data, the network shows better performance on the target tissue samples. Experimental results show that the accuracy of the proposed method is 18.5% higher than that of conventional CNN-based method and the peak intensities of the point spread function (PSF) are more than 20% higher with almost the same training time and processing time. The better compensation performance on target sample could have more advantages when handling complex aberrations, especially the aberrations caused by various histological characteristics, such as refractive index inhomogeneity and biological motion in biological tissues.

Wavefront sensing with optical differentiation powered by deep learning

Deep learning based wavefront sensor for complex wavefront detection in adaptive optical microscopes.

Optical Analog Computing Devices Designed by Deep Neural Network

Deep learning assisted Shack-Hartmann wavefront sensor for direct wavefront detection.

Wavefront Reconstruction Based on Deep Transfer Learning for Microscopy.

Simplifying the detection of optical distortions by machine learning

Fast Wavefront Sensing Method Based on Diffraction Basis Vectors for Tightly Focused Optical Systems

Modal-based Nonlinear Optimization Algorithm for Wavefront Measurement with Under-Sampled Data.

Interferometric Wavefront Sensing System Based on Deep Learning

Fast single pixel modal wavefront sensing using neural networks

Improved Machine Learning Approach for Wavefront Sensing

Wavefront Sensing Based on Phase Contrast Theory and Coherent Optical Processing

Making the unmodulated Pyramid wavefront sensor smart. Closed-loop demonstration of neural network wavefront reconstruction with MagAO-X

High-precision, real-time wavefront sensing via sparse diffractive deep neural networks

Direct wavefront sensing with a plenoptic sensor based on deep learning

Wavefront Sensorless Adaptive Optics Control Algorithm Based on Deep Learning

Single-shot Wavefront Sensing with Nonlocal Thin Film Optical Filters

Experimental Trials With The Optical Differentiation Wavefront Sensor For Extended Objects

High robustness single-shot wavefront sensing method using a near-field profile image and fully-connected retrieval neural network for a high power laser facility

Multi-dimensional wavefront sensing using volumetric meta-optics