Physics-informed and simulation-driven optimization for binary Fourier single-pixel imaging.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Signal Processing

Background:

• Fast Fourier single-pixel imaging (FSI) utilizes binary patterns for rapid image acquisition.
• This speed enhancement often compromises spatial resolution and overall image quality.
• Quantization errors arise from the binary dithering process inherent in FSI.

Purpose of the Study:

• To develop a method for optimizing the trade-off between image quality and speed in FSI.
• To address and compensate for quantization errors in binary-pattern-based single-pixel imaging.
• To enhance the performance of FSI and potentially other single-pixel imaging (SPI) techniques.

Main Methods:

• Proposing and optimizing convolution kernels informed by physical principles and simulation data.
• Developing kernels specifically for both low and high spatial frequencies to counteract dithering effects.
• Validating the proposed method through both computational simulations and experimental setups.

Main Results:

• Demonstrated successful optimization of the image quality-speed trade-off in FSI.
• Successfully compensated for quantization errors using optimized convolution kernels.
• Achieved improved image quality in simulations and experiments compared to standard FSI methods.

Conclusions:

• The proposed method effectively balances imaging speed and quality in FSI.
• Optimized convolution kernels are crucial for mitigating binary dithering artifacts.
• This approach offers a promising advancement for FSI and other SPI modalities.