.png){kind=icon}
.png){kind=icon}
Return
General phase retrieval technology through deep learning
.png){kind=icon}
PRE
AI
DOI：10.1016/j.measurement.2025.118582.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Developed an end-to-end phase retrieval method using AttR2U-Net. • Trained model with Monte Carlo speckle phases for robust mapping. • AttR2U-Net shows higher accuracy and speed than U-Net model. • Enables high-precision phase retrieval for arbitrary images. • Addresses data dependency, generalization in deep learning phase retrieval.
Keywords:
phase retrieval
AttR2U-Net
deep learning
speckle phases
generalization