Block matching low-rank for ghost imaging.

Area of Science:

• Computational imaging
• Optical sensing
• Image reconstruction

Background:

• High-quality ghost imaging (GI) is crucial for scientific and practical applications.
• Reconstructing high-quality images from low sampling data remains a key challenge in GI research.

Purpose of the Study:

• To develop and validate a novel ghost imaging method for high-quality image reconstruction under low sampling conditions.
• To address the limitations of existing GI techniques in low-sampling scenarios.

Main Methods:

• Proposed a method named blocking matching low-rank ghost imaging (BLRGI).
• Leveraged the hypothesis that image nonlocal similar patch matrices exhibit low rank and sparse singular values.
• Employed projected Landweber regularization and blocking matching low-rank denoising techniques.

Main Results:

• Theoretically and experimentally demonstrated the effectiveness of BLRGI.
• BLRGI achieved superior ghost imaging quality compared to 'GI via sparsity constraint,' 'joint iteration GI,' and 'total variation based GI' methods.
• Improvements were observed in peak signal-to-noise ratio, structural similarity index, and visual quality under low sampling.

Conclusions:

• BLRGI offers a significant advancement in ghost imaging, particularly for low-sampling scenarios.
• The method provides a robust approach for obtaining high-fidelity images with reduced data acquisition.
• BLRGI demonstrates potential for enhanced performance in various scientific and practical imaging applications.