Utilization of cross-plane rays for three-dimensional reconstruction by filtered back-projection

Area of Science:

• Medical Imaging
• Computerized Tomography
• Image Reconstruction

Background:

• Conventional computed tomography (CT) algorithms reconstruct objects using ray measurements confined to parallel planes.
• Utilizing only parallel plane data limits the information available for reconstruction, potentially impacting image quality and dose efficiency.

Purpose of the Study:

• To present a novel CT reconstruction algorithm capable of incorporating "cross-plane" rays.
• To demonstrate the theoretical advantages and practical applicability of this enhanced reconstruction method, particularly for emission CT.

Main Methods:

• Developed a reconstruction algorithm that groups ray measurements into 2D projections, filters them, and stores them.
• Implemented back-projection of filtered projections onto a 3D matrix or arbitrary planes.
• Derived expressions for noise power spectrum and reconstruction variance.
• Tested the algorithm using computer-generated data.

Main Results:

• The algorithm effectively reconstructs objects using both parallel and cross-plane rays.
• Noise-to-signal ratio per detected photon is shown to be near the theoretical limit, comparable to conventional CT.
• Demonstrated potential for order-of-magnitude improvements in image quality per unit dose for emission CT.

Conclusions:

• The proposed CT algorithm offers a significant advancement by leveraging cross-plane ray data.
• This method holds particular promise for improving emission CT imaging, enhancing image quality while potentially reducing radiation dose.
• The algorithm achieves near-optimal noise performance, validating its theoretical underpinnings.