Direct PET Image Reconstruction Incorporating Deep Image Prior and a Forward Projection Model

TL;DR

This paper introduces an unsupervised deep image prior framework for direct PET image reconstruction from sinograms, eliminating the need for large training datasets and outperforming traditional algorithms in quality.

Contribution

It presents a novel unsupervised method combining deep image prior and forward projection for PET reconstruction, reducing data requirements and improving image quality.

Findings

Outperforms FBP and ML-EM in PSNR and SSIM metrics

Qualitative improvements in reconstructed image quality

Effective in brain [$^{18}$F]FDG PET simulations

Abstract

Convolutional neural networks (CNNs) have recently achieved remarkable performance in positron emission tomography (PET) image reconstruction. In particular, CNN-based direct PET image reconstruction, which directly generates the reconstructed image from the sinogram, has potential applicability to PET image enhancements because it does not require image reconstruction algorithms, which often produce some artifacts. However, these deep learning-based, direct PET image reconstruction algorithms have the disadvantage that they require a large number of high-quality training datasets. In this study, we propose an unsupervised direct PET image reconstruction method that incorporates a deep image prior framework. Our proposed method incorporates a forward projection model with a loss function to achieve unsupervised direct PET image reconstruction from sinograms. To compare our proposed direct reconstruction method with the filtered back projection (FBP) and maximum likelihood expectation maximization (ML-EM) algorithms, we evaluated using Monte Carlo simulation data of brain [$^{18}$F]FDG PET scans. The results demonstrate that our proposed direct reconstruction quantitatively and qualitatively outperforms the FBP and ML-EM algorithms with respect to peak signal-to-noise ratio and structural similarity index.