Non-invasive accelerated imaging through a scattering medium via multi-stage complexity guidance.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Image Reconstruction

Background:

• Scattering imaging is crucial for various fields, but existing methods struggle to balance non-invasiveness, real-time performance, and image quality.
• Traditional autocorrelation imaging techniques face limitations in speed and clarity.

Purpose of the Study:

• To develop a novel scattering imaging technique that overcomes the limitations of existing methods.
• To enhance the speed, accuracy, clarity, and noise robustness of scattering imaging.

Main Methods:

• Optimization of traditional autocorrelation imaging using multi-stage complexity guidance and an initial acceleration module.
• Introduction of a complexity difference index for effective guidance in phase iterative recovery.
• Implementation of an error-reduction iteration-based initial module for rapid startup.

Main Results:

• The proposed technique significantly accelerates scattering reconstruction speed.
• Achieved higher accuracy and clarity in reconstructed images compared to traditional methods.
• Demonstrated superior robustness to noise than traditional autocorrelation imaging.

Conclusions:

• The novel scattering imaging technique offers a significant advancement in non-invasive, real-time, high-quality imaging.
• The method provides a faster and more robust solution for scattering imaging applications.
• Experimental code is publicly available on GitHub for further research and development.