Repeat and Concatenate: 2D to 3D Image Translation with 3D to 3D Generative Modeling

TL;DR

This paper introduces a simple yet effective 2D to 3D image translation method that concatenates multiple 2D views into a 3D volume and employs 3D generative modeling with neural optimal transport, achieving faithful reconstructions with limited training.

Contribution

The proposed approach simplifies 2D to 3D translation by concatenating views and using 3D generative modeling, avoiding complex alignment and encoding issues.

Findings

Produces high-fidelity 3D reconstructions from 2D views.

Generalizes well across multiple datasets, including out-of-distribution samples.

Requires limited training data for effective results.

Abstract

This paper investigates a 2D to 3D image translation method with a straightforward technique, enabling correlated 2D X-ray to 3D CT-like reconstruction. We observe that existing approaches, which integrate information across multiple 2D views in the latent space, lose valuable signal information during latent encoding. Instead, we simply repeat and concatenate the 2D views into higher-channel 3D volumes and approach the 3D reconstruction challenge as a straightforward 3D to 3D generative modeling problem, sidestepping several complex modeling issues. This method enables the reconstructed 3D volume to retain valuable information from the 2D inputs, which are passed between channel states in a Swin UNETR backbone. Our approach applies neural optimal transport, which is fast and stable to train, effectively integrating signal information across multiple views without the requirement for precise alignment; it produces non-collapsed reconstructions that are highly faithful to the 2D views, even after limited training. We demonstrate correlated results, both qualitatively and quantitatively, having trained our model on a single dataset and evaluated its generalization ability across six datasets, including out-of-distribution samples.