Multi frame holograms batched optimization for binary phase spatial light modulators

Area of Science:

• Computational optics
• Digital holography
• Image processing

Background:

• Traditional phase retrieval methods for computer-generated holograms (e.g., Gerchberg-Saxton, gradient descent) are susceptible to noise and defects.
• Prior work introduced time-averaging of multiple hologram frames (e.g., One-Step Phase-Retrieval, Adaptive One-Step Phase-Retrieval) to mitigate these issues.

Purpose of the Study:

• To develop an advanced phase retrieval technique for generating high-fidelity binary phase holograms.
• To improve algorithmic convergence speed in both Fraunhofer and Fresnel diffraction regimes.
• To offer a robust alternative to existing hologram generation methods.

Main Methods:

• The proposed technique, Multi-Frame Holograms Batched Optimization (MFHBO), utilizes the L-BFGS optimization algorithm.
• MFHBO simultaneously generates a batch of binary phase holograms.
• The method leverages time-averaging of multiple hologram frames for enhanced reconstruction.

Main Results:

• MFHBO demonstrated superior performance in both simulated and experimental comparisons against One-Step Phase-Retrieval and Adaptive One-Step Phase-Retrieval.
• The technique achieved improved fidelity in the average reconstructed image.
• Fast algorithmic convergence was observed in both Fraunhofer and Fresnel regimes.

Conclusions:

• Multi-Frame Holograms Batched Optimization offers a significant advancement in binary phase hologram generation.
• The method provides enhanced image fidelity and faster convergence, outperforming existing techniques.
• The approach is versatile and can be extended to other spatial light modulation applications.