TY - JOUR
T1 - Single-Shot Diffraction Autofocusing
T2 - Distance Prediction via an Untrained Physics-Enhanced Network
AU - Tang, Ju
AU - Wu, Ji
AU - Zhang, Jiawei
AU - Ren, Zhenbo
AU - Di, Jianglei
AU - Zhao, Jianlin
N1 - Publisher Copyright:
© 2009-2012 IEEE.
PY - 2022/2/1
Y1 - 2022/2/1
N2 - Deep learning technology has shown excellent performances and successful applications in optical information processing. However, the long-time training, large amount of manually labeled data and generalization capability hinder the application of deep neural network (DNN) under supervised learning. The deep image prior (DIP) opinion promotes the development of untrained neural network, which can learn from one image. Here we propose a DIP-based strategy to nest the DNN into a physical model for finding the optimal solution in a univariate optimization problem. The untrained physics-enhanced network (UPN) is proposed to predict the diffraction distance via only one diffraction pattern of a known phase object. Simulation and experimental results show that the UPN can be used to predict the distance precisely and consistently with different targets, diffraction distances as well as phase ranges, while it only takes a little time for training. In addition, the trained UPN can generalize to the other targets as long as the actual diffraction process keeps the same. Compared with the autofocusing metrics of holographic reconstruction and traversal method, the UPN has advantages in speed and accuracy, and it also has good noise resistance, which are all meaningful for the autofocusing of holographic reconstruction and imaging.
AB - Deep learning technology has shown excellent performances and successful applications in optical information processing. However, the long-time training, large amount of manually labeled data and generalization capability hinder the application of deep neural network (DNN) under supervised learning. The deep image prior (DIP) opinion promotes the development of untrained neural network, which can learn from one image. Here we propose a DIP-based strategy to nest the DNN into a physical model for finding the optimal solution in a univariate optimization problem. The untrained physics-enhanced network (UPN) is proposed to predict the diffraction distance via only one diffraction pattern of a known phase object. Simulation and experimental results show that the UPN can be used to predict the distance precisely and consistently with different targets, diffraction distances as well as phase ranges, while it only takes a little time for training. In addition, the trained UPN can generalize to the other targets as long as the actual diffraction process keeps the same. Compared with the autofocusing metrics of holographic reconstruction and traversal method, the UPN has advantages in speed and accuracy, and it also has good noise resistance, which are all meaningful for the autofocusing of holographic reconstruction and imaging.
KW - Deep image prior
KW - physical model
KW - prediction of diffraction distance
KW - untrained physics-enhanced network
UR - http://www.scopus.com/inward/record.url?scp=85122293197&partnerID=8YFLogxK
U2 - 10.1109/JPHOT.2021.3138548
DO - 10.1109/JPHOT.2021.3138548
M3 - 文章
AN - SCOPUS:85122293197
SN - 1943-0655
VL - 14
JO - IEEE Photonics Journal
JF - IEEE Photonics Journal
IS - 1
ER -