[A new Bayesian reconstruction algorithm for PET images based on correction of the detected sinogram data]

Area of Science:

• Medical Imaging
• Computational Science
• Image Reconstruction

Context:

• Positron-emission tomography (PET) imaging is crucial for medical diagnosis.
• Traditional Bayesian reconstruction methods can be limited by noise in PET data.
• Improving the quality of reconstructed PET images is essential for accurate analysis.

Purpose:

• To develop a novel coupled feedback (CF) iterative model for enhanced Bayesian reconstruction of PET images.
• To address the challenge of noisy sinogram data in PET image reconstruction.
• To improve the performance of PET image reconstruction algorithms.

Summary:

• A coupled feedback (CF) iterative algorithm was developed and applied to update noisy PET sinogram data using the latest reconstructed image.
• The method incorporates linear filtering, Wiener filtering, and projection of reconstructed images.
• Simulated phantom data was used to evaluate the CF iterative algorithm.

Impact:

• The CF iterative model demonstrated superior performance compared to traditional Bayesian reconstruction methods in visual and quantitative comparisons.
• The proposed sinogram-correcting method effectively improves original measurement data.
• This approach is applicable to PET image reconstruction and other high-noise reconstruction tasks.