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.

Deep Learning Microscopy

Deep Learning Enhanced Fluorescence Emission Difference Microscopy

Deep-Learning-Based Image Reconstruction and Enhancement in Optical Microscopy

Universal and High-Fidelity Resolution Extending for Fluorescence Microscopy Using a Single-Training Physics-Informed Sparse Neural Network

Wavefront Reconstruction Based on Deep Transfer Learning for Microscopy.

Roadmap on Deep Learning for Microscopy

Deep-learning Light-Field Microscopy with Improved Resolution and Reconstruction Speed

Enhancing fluorescence microscopy performance with Richardson-Lucy based deep learning

Toward a Thinking Microscope: Deep Learning in Optical Microscopy and Image Reconstruction

Real-Time Resolution Enhancement of Confocal Laser Scanning Microscopy via Deep Learning

High-Throughput Deep Learning Microscopy Using Multi-Angle Super-Resolution

Deep learning enhanced mobile-phone microscopy

Super-Resolution and Large Depth of Field Model for Optical Microscope Imaging

Deep-learning-augmented Microscopy for Super-Resolution Imaging of Nanoparticles.

A Survey on Applications of Deep Learning in Microscopy Image Analysis

Incorporating the Image Formation Process into Deep Learning Improves Network Performance

Large depth-of-field ultra-compact microscope by progressive optimization and deep learning

Advances in Deep Learning for Super-Resolution Microscopy(Invited)

Deep learning enables cross-modality super-resolution in fluorescence microscopy

Deep-learning-enhanced light-field microscopy

Zero-shot learning enables instant denoising and super-resolution in optical fluorescence microscopy