NAF: Neural Attenuation Fields for Sparse-View CBCT Reconstruction

Abstract

This paper proposes a novel and fast self-supervised solution for sparse-view CBCT reconstruction (Cone Beam Computed Tomography) that requires no external training data. Specifically, the desired attenuation coefficients are represented as a continuous function of 3D spatial coordinates, parameterized by a fully-connected deep neural network. We synthesize projections discretely and train the network by minimizing the error between real and synthesized projections. A learning-based encoder entailing hash coding is adopted to help the network capture high-frequency details. This encoder outperforms the commonly used frequency-domain encoder in terms of having higher performance and efficiency, because it exploits the smoothness and sparsity of human organs. Experiments have been conducted on both human organ and phantom datasets. The proposed method achieves state-of-the-art accuracy and spends reasonably short computation time.

arXiv commentMICCAI2022 (Oral)

References in corpus (4)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Instant Neural Graphics Primitives with a Multiresolution Hash Encoding
• Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains
• Learned Primal-dual Reconstruction

Cited by in corpus (7)

• 3DGR-CT: Sparse-View CT Reconstruction with a 3D Gaussian Representation
• Geometry-Aware Attenuation Learning for Sparse-View CBCT Reconstruction
• Continuous 3D Myocardial Motion Tracking via Echocardiography
• NeCA: 3D Coronary Artery Tree Reconstruction from Two 2D Projections via Neural Implicit Representation
• GuidedRec: Guiding Ill-Posed Unsupervised Volumetric Recovery
• Equivariance2Inverse: A Practical Self-Supervised CT Reconstruction Method Benchmarked on Real, Limited-Angle, and Blurred Data
• Fit Pixels, Get Labels: Meta-learned Implicit Networks for Image Segmentation