Single-image Tomography: 3D Volumes from 2D Cranial X-Rays

TL;DR

This paper introduces a deep learning method for reconstructing 3D volumes from 2D X-ray images, using a coarse-to-fine approach trained on a large simulated dataset, enabling applications like re-rendering and pose changes.

Contribution

It presents a novel deep learning framework for single-image 3D reconstruction from X-rays, including a new dataset and a coarse-to-fine volume generation strategy.

Findings

High accuracy in 3D volume reconstruction from X-rays.

Effective application to real X-ray images and re-rendering tasks.

Outperforms previous methods in evaluation and user studies.

Abstract

As many different 3D volumes could produce the same 2D x-ray image, inverting this process is challenging. We show that recent deep learning-based convolutional neural networks can solve this task. As the main challenge in learning is the sheer amount of data created when extending the 2D image into a 3D volume, we suggest firstly to learn a coarse, fixed-resolution volume which is then fused in a second step with the input x-ray into a high-resolution volume. To train and validate our approach we introduce a new dataset that comprises of close to half a million computer-simulated 2D x-ray images of 3D volumes scanned from 175 mammalian species. Applications of our approach include stereoscopic rendering of legacy x-ray images, re-rendering of x-rays including changes of illumination, view pose or geometry. Our evaluation includes comparison to previous tomography work, previous learning methods using our data, a user study and application to a set of real x-rays.