Single-shot coherent diffraction imaging with weighted complex structure tensor

Area of Science:

• Coherent imaging
• Computational imaging
• Image reconstruction

Background:

• Phase retrieval is essential for reconstructing amplitude and phase in coherent imaging.
• Single-shot phase retrieval faces challenges due to limited measurements, hindering fine detail resolution.

Purpose of the Study:

• To develop an unsupervised, noise-robust single-shot phase retrieval algorithm.
• To enhance the reconstruction of fine details and texture in complex-valued objects.

Main Methods:

• Integration of alternating direction optimization.
• Application of weighted complex structure tensor total variation regularization.
• Balancing data fitting with texture preservation.

Main Results:

• The proposed algorithm demonstrates noise robustness.
• Successful reconstruction of complex amplitude for texture-rich objects.
• Enhanced fidelity and fine detail resolution in simulations and experiments.

Conclusions:

• The developed method effectively addresses limitations in single-shot phase retrieval.
• It offers a robust solution for reconstructing complex-valued objects with high fidelity.
• The approach shows significant potential for advanced imaging applications.