Filtered Back-Projection Reconstruction with Non-Uniformly Under-Sampled Projections

Area of Science:

• Medical Imaging
• Image Reconstruction
• Computational Imaging

Background:

• Tomographic imaging systems typically require uniform angular sampling for accurate image reconstruction.
• Non-uniform or under-sampled data can lead to aliasing artifacts, degrading image quality.
• Existing methods may struggle with reconstructing images from non-uniformly sampled sinograms.

Purpose of the Study:

• To address challenges in tomographic image reconstruction with non-uniformly under-sampled sinograms.
• To propose a novel closed-form formula for converting non-uniform sinograms to uniform ones.
• To demonstrate the efficacy of the proposed method using the filtered back-projection algorithm.

Main Methods:

• A case study involving a non-uniformly under-sampled sinogram was analyzed.
• A closed-form formula was developed to transform non-uniform angular sampling to uniform sampling.
• The filtered back-projection (FBP) algorithm was applied to the processed sinogram for image reconstruction.

Main Results:

• The proposed formula enables the conversion of non-uniformly sampled sinograms to uniformly sampled sinograms.
• Image reconstruction was successfully performed using the filtered back-projection algorithm on the converted data.
• The method provides an exact conversion under the assumption of a band-limited sinogram.

Conclusions:

• The developed formula offers a viable solution for reconstructing images from non-uniformly under-sampled tomographic data.
• This approach can potentially mitigate aliasing artifacts and improve image quality in specific tomographic imaging scenarios.
• Further research may be needed to address the limitations of the band-limited assumption in real-world applications.