Coarse-to-fine golden section autofocus for ptychography using gradient and frequency metrics

Area of Science:

• Optics and Photonics
• Computational Imaging

Background:

• Accurate sample-to-detector distance is critical for high-quality ptychographic reconstruction.
• Errors in this parameter lead to phase distortions and defocus artifacts, degrading image quality.

Purpose of the Study:

• To develop a robust and efficient autofocus strategy for ptychographic imaging.
• To improve focusing accuracy and convergence speed while minimizing computational cost.

Main Methods:

• A two-stage autofocus strategy utilizing the golden section search algorithm.
• Combines Tenengrad gradient metric for coarse localization and band-limited frequency energy (BLFE) for fine refinement.
• Exploits spectrally constrained image sharpness for effective focusing.

Main Results:

• The proposed method demonstrates superior convergence speed and focusing accuracy compared to conventional algorithms.
• Effectively eliminates defocus errors, leading to higher spatial resolution and improved image contrast.
• Achieves superior quantitative metrics with a limited computational budget.

Conclusions:

• The developed hybrid autofocus strategy provides a practical solution for high-quality ptychographic imaging.
• Enhances image fidelity and quantitative performance in ptychography.
• Offers an efficient approach to address defocus errors in imaging systems.