Improving PSNR and computational efficiency in orthogonal ghost imaging.

Area of Science:

• Optics and Photonics
• Computational Imaging
• Signal Processing

Background:

• Traditional ghost imaging often relies on random or sinusoidal patterns, which can limit reconstruction fidelity and efficiency.
• Environmental noise can degrade the performance of existing ghost imaging systems.
• There is a need for improved ghost imaging techniques that offer higher quality, faster reconstruction, and reduced computational cost.

Purpose of the Study:

• To introduce a novel orthogonal ghost imaging (OGI) method utilizing two-dimensional discrete cosine transform (2D-DCT) patterns.
• To enhance image quality, reconstruction speed, and computational efficiency compared to existing ghost imaging approaches.
• To demonstrate the effectiveness of OGI with reduced sampling rates and noise reduction techniques.

Main Methods:

• Development of an orthogonal ghost imaging (OGI) method employing structured two-dimensional discrete cosine transform (2D-DCT) patterns.
• Derivation of a new reconstruction formula tailored for the OGI method.
• Implementation of a phase-shift illumination pattern technique for environmental noise reduction.

Main Results:

• The proposed OGI method demonstrates superior reconstruction fidelity and computational efficiency over random and sinusoidal pattern-based ghost imaging.
• High-quality image reconstruction is achieved even with significantly reduced sampling rates (e.g., 30% of measurements meeting Shannon entropy predictions).
• The OGI technique consistently outperforms differential and sinusoidal ghost imaging in signal-to-noise ratio (SNR) and reconstruction efficiency.

Conclusions:

• The orthogonal ghost imaging (OGI) method based on 2D-DCT patterns offers a significant advancement in ghost imaging technology.
• The technique provides a promising direction for developing efficient, high-fidelity, and low-cost ghost imaging systems.
• The OGI method effectively addresses limitations of traditional approaches, particularly in terms of speed, quality, and data requirements.