hutao568/illumination-coding-meets-uncertainty-learning

Project repository for deep learning coded FPM with uncertainty quantification

Reliable deep learning-based phase imaging with uncertainty quantification

Python (Keras) implementation of paper: Reliable deep learning-based phase imaging with uncertainty quantification. We provide model, pre-trained weight, test data and a quick demo.

Citation

If you find this project useful in your research, please consider citing our paper:

**Yujia Xue, Shiyi Cheng, Yunzhe Li, and Lei Tian, "Reliable deep-learning-based phase imaging with uncertainty quantification," Optica 6, 618-629 (2019)

Abstract

Emerging deep-learning (DL)-based techniques have significant potential to revolutionize biomedical imaging. However, one outstanding challenge is the lack of reliability assessment in the DL predictions, whose errors are commonly revealed only in hindsight. Here, we propose a new Bayesian convolutional neural network (BNN)-based framework that overcomes this issue by quantifying the uncertainty of DL predictions. Foremost, we show that BNN-predicted uncertainty maps provide surrogate estimates of the true error from the network model and measurement itself. The uncertainty maps characterize imperfections often unknown in real-world applications, such as noise, model error, incomplete training data, and out-of-distribution testing data. Quantifying this uncertainty provides a per-pixel estimate of the confidence level of the DL prediction as well as the quality of the model and data set. We demonstrate this framework in the application of large space–bandwidth product phase imaging using a physics-guided coded illumination scheme. From only five multiplexed illumination measurements, our BNN predicts gigapixel phase images in both static and dynamic biological samples with quantitative credibility assessment. Furthermore, we show that low-certainty regions can identify spatially and temporally rare biological phenomena. We believe our uncertainty learning framework is widely applicable to many DL-based biomedical imaging techniques for assessing the reliability of DL predictions.

Requirements

python 3.6

keras 2.1.2

tensorflow 1.4.0

numpy 1.14.3

h5py 2.7.1

matplotlib 2.1.2

Phase prediction and Uncertainty quantification Framework

Network Structure

How to run the demo

Make sure you have all dependent packages installed correctly and run demo.py.

Results

On SEEN cell type

On UNSEEN cell type

Uncertainty learning framework identifies spatially and temporally rare biological phenomena.

License

This project is licensed under the terms of the BSD-3-Clause license. see the LICENSE file for details