Enhancing Dynamic CT Image Reconstruction with Neural Fields and Optical Flow

2024-06-03

AI Summary

Dynamic CT reconstruction faces a spatiotemporal sampling imbalance: high temporal resolution necessitates severe spatial undersampling, leading to motion artifacts and temporal inconsistency. To address this, we propose a physics-guided reconstruction method that jointly leverages neural field representation and optical flow equation constraints. Specifically, we introduce, for the first time, a PDE-based optical flow regularizer into the neural field dynamic inverse problem optimization framework, explicitly modeling continuous organ motion and overcoming the limitations of purely data-fidelity-driven approaches. Our method employs neural fields for compact, high-dimensional spatiotemporal signal representation, enforces kinematic plausibility via the optical flow constraint, and performs end-to-end optimization guided by PSNR. Experiments demonstrate significant PSNR improvement over unregularized neural fields and superior reconstruction quality compared to conventional gridded solvers under identical undersampling conditions.

Technology Category

• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems
• Computer Vision: Low Level & Physics-based Vision
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

In this paper, we investigate image reconstruction for dynamic Computed Tomography. The motion of the target with respect to the measurement acquisition rate leads to highly resolved in time but highly undersampled in space measurements. Such problems pose a major challenge: not accounting for the dynamics of the process leads to a poor reconstruction with non-realistic motion. Variational approaches that penalize time evolution have been proposed to relate subsequent frames and improve image quality based on classical grid-based discretizations. Neural fields have emerged as a novel way to parameterize the quantity of interest using a neural network with a low-dimensional input, benefiting from being lightweight, continuous, and biased towards smooth representations. The latter property has been exploited when solving dynamic inverse problems with neural fields by minimizing a data-fidelity term only. We investigate and show the benefits of introducing explicit motion regularizers for dynamic inverse problems based on partial differential equations, namely, the optical flow equation, for the optimization of neural fields. We compare it against its unregularized counterpart and show the improvements in the reconstruction. We also compare neural fields against a grid-based solver and show that the former outperforms the latter in terms of PSNR in this task.

Problem

• Improving dynamic CT image reconstruction accuracy
• Addressing motion artifacts in undersampled spatial measurements
• Enhancing neural fields with optical flow regularization

Innovation

• Neural fields parameterize dynamic CT reconstruction
• Optical flow regularizes motion in neural fields
• Neural fields outperform grid-based solvers in PSNR