Lensless extended depth of field imaging using PSF correction and pre-denoising

Area of Science:

• Computational Imaging
• Optical Engineering
• Image Processing

Background:

• Lensless cameras offer compact designs for extended depth-of-field (EDoF).
• Existing EDoF lensless cameras compromise diffraction efficiency for depth-invariant point spread functions (PSFs).
• This compromise leads to PSFs with long tails, narrow modulation transfer functions, and noise amplification during deconvolution.

Purpose of the Study:

• To develop a novel method for improving image quality in lensless EDoF cameras.
• To overcome the limitations of low diffraction efficiency and noise amplification in current lensless imaging systems.
• To achieve high-quality imaging across a large depth-of-field range in a single exposure.

Main Methods:

• A two-measurement PSF correction method to enhance the accuracy of measured PSFs.
• A pre-denoising strategy incorporating a denoising network (DN).
• A joint training framework enabling the reconstruction algorithm to guide the DN's convergence.

Main Results:

• Accurate PSF measurement and effective noise reduction were achieved.
• High-quality imaging performance was demonstrated across a wide range of depths.
• The proposed methods significantly improved imaging quality without sacrificing diffraction efficiency.

Conclusions:

• The developed PSF correction and pre-denoising strategy effectively addresses image quality degradation in lensless EDoF cameras.
• This approach overcomes the trade-off between PSF consistency and diffraction efficiency.
• The method offers a computationally efficient and time-saving solution for high-quality single-exposure EDoF imaging.