Phase Extraction Neural Network (PhENN) and Its Low-photon Application
PhENN was trained to solve end-to-end inverse problems in lensless computational imaging. It was also applied to solve the low-photon phase retrieval problem.
Sinha et al., Optica (2017)
Goy et al., Physical Review Letters (2018)
Limited-angle Tomography using Deep Learning
The number of measurements needed for tomographic reconstruction was largely reduced with the use of Deep Neural Network. Therefore, high-resolution 3D reconstruction was possible with only 20 degree of scanning.
Goy et al., PNAS (2019)
Novel DNN Architecture Design for Robust Low-photon Phase Retrieval
In the low-photon condition, measurements suffer from shot noise, degrading the image fidelity of reconstructions. This novel DNN architecture deals with robustness in low-photon phase retrieval.
Deng et al., Light: Science and Applications (2020)