High-quality phase-only Fourier hologram generation with camera-in-the-loop.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Digital Holography

Background:

• Phase-only Fourier holograms are crucial for optical reconstruction.
• Traditional camera-in-the-loop (CITL) methods optimize phase holograms, which can be computationally intensive.
• Achieving high-quality optical reconstruction with phase-only holograms remains a challenge.

Purpose of the Study:

• To propose an optimization method for Fourier holograms enabling high-quality optical reconstruction of phase-only Fourier holograms.
• To introduce a novel camera-in-the-loop (CITL) framework that optimizes input amplitudes for superior hologram quality.
• To develop a neural network, PoFNet, for efficient inference of optimized input amplitudes, addressing the limitations of traditional CITL algorithms.

Main Methods:

• Defining input image amplitude as an optimization plane within a CITL framework.
• Optimizing input amplitudes for high correlation with original images in Fourier holograms.
• Developing and training a neural network (PoFNet) using the proposed CITL for non-ideal optical propagation.

Main Results:

• The proposed CITL method and PoFNet effectively reduce noise in the Fourier optical path.
• Optical experiments validated the superior optical reconstruction quality achieved by optimizing input amplitudes.
• PoFNet significantly reduces the time-consuming nature of traditional CITL by inferring optimized amplitudes.

Conclusions:

• The novel optimization strategy for Fourier holograms enhances optical reconstruction quality.
• The integration of PoFNet with the proposed CITL offers an efficient and effective solution for generating high-quality phase-only holograms.
• This research contributes to advancements in digital holography and optical imaging systems.