Image Reconstruction with Maclaurin Series Expansion.

Area of Science:

• Medical Imaging
• Computational Imaging
• Image Reconstruction

Background:

• Current imaging systems often require extensive data for accurate image reconstruction.
• Prior knowledge or training data is typically necessary for robust reconstruction, especially with limited measurements.
• Theoretical frameworks are crucial for advancing the fundamental understanding of image reconstruction.

Purpose of the Study:

• To investigate a novel theoretical approach for image reconstruction using minimal data.
• To explore the feasibility of reconstructing a complete dataset from a small scanning angular range.
• To develop a method for trustworthy reconstruction without relying on prior knowledge or training data.

Main Methods:

• Developed a Maclaurin series expansion in the Fourier domain under idealized conditions (no noise, continuous signals, perfect computation).
• Demonstrated the convergence of this expansion across the entire Fourier space.
• Utilized computer simulations to illustrate 2D image reconstruction from a truncated Fourier-domain Maclaurin series expansion.

Main Results:

• Showed that a Maclaurin series expansion in the Fourier domain can yield a complete dataset from limited angular measurements.
• Confirmed the convergence of the expansion, theoretically enabling full Fourier space coverage.
• Successfully reconstructed a 2D spatial-domain image using a truncated expansion, validating the theoretical approach.

Conclusions:

• The theoretical framework supports the potential for high-fidelity image reconstruction using significantly reduced measurement data.
• This approach offers a pathway towards data-efficient imaging without prior information or machine learning models.
• While currently theoretical, the findings provide a foundation for future research into practical, low-data imaging systems.