Learning Perspective Deformation in X-Ray Transmission Imaging

TL;DR

This paper introduces a framework using dual-view X-ray images to correct perspective deformation, improving geometric accuracy in medical imaging with deep learning models that are robust to real-world challenges.

Contribution

It formulates a novel dual-view approach for perspective correction and evaluates two deep learning networks, demonstrating improved accuracy and robustness in various imaging scenarios.

Findings

Complementary views outperform orthogonal or single views.

Pix2pixGAN excels in polar space correction.

TransU-Net achieves comparable performance in Cartesian space.

Abstract

In cone-beam X-ray transmission imaging, perspective deformation causes difficulty in direct, accurate geometric assessments of anatomical structures. In this work, the perspective deformation correction problem is formulated and addressed in a framework using two complementary (180{\deg}) views. The complementary view setting provides a practical way to identify perspectively deformed structures by assessing the deviation between the two views. It also provides bounding information and reduces uncertainty for learning perspective deformation. Two representative networks Pix2pixGAN and TransU-Net for correcting perspective deformation are investigated. Experiments on numerical bead phantom data demonstrate the advantage of complementary views over orthogonal views or a single view. They show that Pix2pixGAN as a fully convolutional network achieves better performance in polar space than Cartesian space, while TransU-Net as a transformer-based hybrid network achieves comparable performance in Cartesian space to polar space. Further study demonstrates that the trained model has certain tolerance to geometric inaccuracy within calibration accuracy. The efficacy of the proposed framework on synthetic projection images from patients' chest and head data as well as real cadaver CBCT projection data and its robustness in the presence of bulky metal implants and surgical screws indicate the promising aspects of future real applications.