CFConv: complex-valued Fourier convolutional layers for artifact suppression in cylindrical holographic...

Area of Science:

• Optics and Photonics
• Computer Vision
• Digital Imaging

Background:

• Cylindrical holography enables 360° full-view displays but faces challenges due to asymmetric diffraction models.
• Asymmetry causes high-frequency information loss and energy attenuation, leading to artifacts and reduced image quality in reconstructions.

Purpose of the Study:

• To address the limitations of existing optimization methods in cylindrical holography.
• To propose a novel approach for compensating frequency-domain imbalance in holographic reconstruction.

Main Methods:

• Introduction of complex-valued Fourier convolutional layers tailored for holographic diffraction processes.
• Integration of these layers into the phase prediction stage to preserve high-frequency information.

Main Results:

• The proposed complex Fourier convolutional layers effectively preserve high-frequency information around object contours.
• Suppression of reconstruction artifacts and significant improvement in the fidelity and sharpness of reconstructed images.
• Demonstrated enhanced image quality in cylindrical computer-generated holography.

Conclusions:

• Complex-valued Fourier convolutional layers offer a robust solution for frequency-domain imbalance in cylindrical holography.
• The method enhances image fidelity and sharpness, paving the way for higher-quality 360° holographic displays.
• This advancement addresses key challenges in achieving artifact-free holographic reconstructions.