MORE: Multi-Organ Medical Image REconstruction Dataset

TL;DR

MORE addresses the generalization gap in CT reconstruction by providing a large, multi-organ dataset with 9 anatomies and 15 lesion types, enabling robust training and benchmarking. The authors introduce GIFT, a Gaussian-based optimization framework that represents the 3D volume as a sum of Gaussians and optimizes in the measurement domain, formalized as $x^* = \arg\min_x \mathcal{E}(Ax,y) + R(x)$ with $A$ the Radon operator. Across 60–180 view sparse-view benchmarks, GIFT and other optimization-based methods show stronger generalization to unseen anatomies and lesions than pretrained DL models, and training on MORE improves cross-dataset generalization (e.g., to AAPM-Mayo Abdomen and foot-fracture cases). The work provides a publicly accessible resource and code to advance robust, clinically relevant CT reconstruction in heterogeneous settings.

Abstract

CT reconstruction provides radiologists with images for diagnosis and treatment, yet current deep learning methods are typically limited to specific anatomies and datasets, hindering generalization ability to unseen anatomies and lesions. To address this, we introduce the Multi-Organ medical image REconstruction (MORE) dataset, comprising CT scans across 9 diverse anatomies with 15 lesion types. This dataset serves two key purposes: (1) enabling robust training of deep learning models on extensive, heterogeneous data, and (2) facilitating rigorous evaluation of model generalization for CT reconstruction. We further establish a strong baseline solution that outperforms prior approaches under these challenging conditions. Our results demonstrate that: (1) a comprehensive dataset helps improve the generalization capability of models, and (2) optimization-based methods offer enhanced robustness for unseen anatomies. The MORE dataset is freely accessible under CC-BY-NC 4.0 at our project page https://more-med.github.io/

Figures (3)

• Figure 1: Data distribution of the MORE CT part, containing 9 anatomies and 15 lesion types.
• Figure 2: Framework of our baseline solution GIFT.
• Figure 3: Performance of 60-view SV-CT on 15 types of lesions within the MORE dataset.