Improving the performance of ghost imaging via measurement-driven framework.

Area of Science:

• Computational imaging
• Optical sensing
• Image processing

Background:

• High-quality reconstruction in ghost imaging is crucial, especially under low sampling rates.
• Achieving precise imaging results with limited data is a significant challenge and active research area in ghost imaging.

Purpose of the Study:

• To introduce an optimization scheme for speckle patterns in ghost imaging to enhance reconstruction quality.
• To improve ghost imaging performance using a measurement-driven framework inspired by compressed sensing matrix optimization.

Main Methods:

• An optimization scheme for speckle patterns using a measurement-driven framework.
• Alternating optimization of the sampling matrix and sparse basis using sparse coefficients from low-dimension pseudo-measurements.
• Analytical solutions for the optimization process, followed by non-negative constraint and binary quantization of the sampling matrix.

Main Results:

• The proposed scheme demonstrates superior reconstruction quality compared to existing methods at low sampling rates.
• Achieved significant improvements in peak signal-to-noise ratio (PSNR) and mean structural similarity index (MSSIM).
• Effective ghost imaging reconstruction was demonstrated at a sampling rate as low as 6.5%, yielding an MSSIM of 0.787 and PSNR of 17.078 dB.

Conclusions:

• The developed measurement-driven framework and optimized speckle patterns significantly enhance ghost imaging reconstruction quality under low sampling conditions.
• The proposed method offers a robust solution for achieving high-fidelity ghost imaging with reduced data acquisition.
• This approach represents a notable advancement in ghost imaging technology, particularly for applications requiring efficient data usage.