Deep learning enhanced quantum holography with undetected photons.

Area of Science:

• Optics and Photonics
• Quantum Information Science
• Artificial Intelligence in Imaging

Background:

• Holography is crucial for 3D imaging, with quantum holography with undetected photons (QHUP) enabling complex amplitude image capture.
• Practical QHUP is hindered by phase disturbances, low visibility, and limited spatial resolution.
• Deep learning (DL) offers powerful data processing capabilities to address these QHUP limitations.

Purpose of the Study:

• To develop and demonstrate a deep learning-based approach to improve QHUP.
• To extract high-quality images from single-shot holograms using DL.
• To enhance spatial resolution, reduce noise and distortion, and improve imaging speed and resilience.

Main Methods:

• Implementation of a novel deep learning QHUP (DL-QHUP) methodology.
• Utilizing DL algorithms for image extraction from single-shot holograms.
• Demonstration of DL-QHUP for advanced holographic imaging.

Main Results:

• Achieved significantly reduced noise and distortion in extracted images.
• Demonstrated notable enhancement in spatial resolution.
• Enabled high-speed imaging with superior noise resilience.

Conclusions:

• DL-QHUP provides a transformative solution for overcoming key QHUP limitations.
• The DL-QHUP methodology offers improved performance for diverse applications, including biomedical imaging and remote sensing.
• This advancement represents a significant leap in holography, unlocking new imaging possibilities in challenging environments.