Multibranch Generative Models for Multichannel Imaging with an Application to PET/CT Synergistic Reconstruction

TL;DR

This paper introduces a multibranch generative model based on VAEs for synergistic reconstruction of medical images, improving quality in low-dose PET/CT imaging by learning from image pairs and incorporating a regularizer.

Contribution

The paper proposes a novel multibranch VAE-based framework for synergistic medical image reconstruction, integrating learned models as regularizers to enhance image quality.

Findings

Improved image quality in low-dose PET/CT reconstructions.

Effective denoising demonstrated on MNIST and PET/CT datasets.

Potential for generative models to advance medical imaging reconstruction.

Abstract

This paper presents a novel approach for learned synergistic reconstruction of medical images using multibranch generative models. Leveraging variational autoencoders (VAEs), our model learns from pairs of images simultaneously, enabling effective denoising and reconstruction. Synergistic image reconstruction is achieved by incorporating the trained models in a regularizer that evaluates the distance between the images and the model. We demonstrate the efficacy of our approach on both Modified National Institute of Standards and Technology (MNIST) and positron emission tomography (PET)/computed tomography (CT) datasets, showcasing improved image quality for low-dose imaging. Despite challenges such as patch decomposition and model limitations, our results underscore the potential of generative models for enhancing medical imaging reconstruction.