LoDoPaB-CT, a benchmark dataset for low-dose computed tomography reconstruction

Leuschner, Johannes; Schmidt, Maximilian; Baguer, Daniel Otero; Maaß, Peter

doi:10.1038/s41597-021-00893-z

Cited by 66 publications

(47 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our experiments, we use the LoDoPaB-CT dataset [62] to replicate the challenges that arise from low-dose CT measurements. The dataset contains over 40,000 normal-dose, medical CT images from the human thorax from around 800 patients.…”

Section: Computed Tomographymentioning

confidence: 99%

“…Since there are no real groundtruth images available, high-quality reconstructions are used for training. For LoDoPaB-CT, reconstruction from normal-dose CT measurements and for fastMRI reconstruction from fully sampled MRI measurements are used instead [62,65]. These reconstructions are not free of noise, so we use an additional dequantization step and add random Gaussian noise in the order of the background noise to the training images.…”

Section: Base Distributionmentioning

confidence: 99%

See 1 more Smart Citation

Conditional Invertible Neural Networks for Medical Imaging

et al. 2021

Self Cite

View full text Add to dashboard Cite

Over recent years, deep learning methods have become an increasingly popular choice for solving tasks from the field of inverse problems. Many of these new data-driven methods have produced impressive results, although most only give point estimates for the reconstruction. However, especially in the analysis of ill-posed inverse problems, the study of uncertainties is essential. In our work, we apply generative flow-based models based on invertible neural networks to two challenging medical imaging tasks, i.e., low-dose computed tomography and accelerated medical resonance imaging. We test different architectures of invertible neural networks and provide extensive ablation studies. In most applications, a standard Gaussian is used as the base distribution for a flow-based model. Our results show that the choice of a radial distribution can improve the quality of reconstructions.

show abstract

Section: Computed Tomographymentioning

confidence: 99%

Section: Base Distributionmentioning

confidence: 99%

Conditional Invertible Neural Networks for Medical Imaging

et al. 2021

Self Cite

View full text Add to dashboard Cite

show abstract

“…In our experiments, we use the LoDoPaB-CT dataset [55] to replicate the challenges that arise from low-dose CT measurements. The dataset contains over 40 000 normal-dose, medical CT images from the human thorax from around 800 patients.…”

Section: Computed Tomographymentioning

confidence: 99%

“…Since there are no real ground truth images available, high-quality reconstructions are used for training. For LoDoPaB-CT, reconstruction from normal-dose CT measurements and for fastMRI reconstruction from fully sampled MRI measurements are used instead [55,58]. These reconstructions are not free of noise, so we use an additional dequantization step and add random Gaussian noise in the order of the background noise to the training images.…”

Section: Training With Additional Noisementioning

confidence: 99%

Conditional Invertible Neural Networks for Medical Imaging

Denker¹,

Schmidt²,

Leuschner³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Over the last years, deep learning methods have become an increasingly popular choice to solve tasks from the field of inverse problems. Many of these new data-driven methods have produced impressive results, although most only give point estimates for the reconstruction. However, especially in the analysis of ill-posed inverse problems, the study of uncertainties is essential. In our work, we apply generative flow-based models based on invertible neural networks to two challenging medical imaging tasks, i.e. low-dose computed tomography and accelerated medical resonance imaging. We test different architectures of invertible neural networks and provide extensive ablation studies. In most applications, a standard Gaussian is used as the base distribution for a flow-based model. Our results show that the choice of a radial distribution can improve the quality of reconstructions.

show abstract

“…Here max x is the maximal possible pixel value of the image and in general we have max x = 1. Nevertheless, in [41] it is proposed to choose max x = max(x) − min(x) for CT since here the pixel values are far from most common values. Herewith we avoid too optimistic results.…”

mentioning

confidence: 99%

PatchNR: Learning from Small Data by Patch Normalizing Flow Regularization

Altekrüger¹,

Denker²,

Hagemann³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Learning neural networks using only a small amount of data is an important ongoing research topic with tremendous potential for applications. In this paper, we introduce a regularizer for the variational modeling of inverse problems in imaging based on normalizing flows. Our regularizer, called patchNR, involves a normalizing flow learned on patches of very few images. The subsequent reconstruction method is completely unsupervised and the same regularizer can be used for different forward operators acting on the same class of images. By investigating the distribution of patches versus those of the whole image class, we prove that our variational model is indeed a MAP approach. Our model can be generalized to conditional patchNRs, if additional supervised information is available. Numerical examples for low-dose CT, limited-angle CT and superresolution of material images demonstrate that our method provides high quality results among unsupervised methods, but requires only few data.

show abstract

LoDoPaB-CT, a benchmark dataset for low-dose computed tomography reconstruction

Cited by 66 publications

References 46 publications

Conditional Invertible Neural Networks for Medical Imaging

Conditional Invertible Neural Networks for Medical Imaging

Conditional Invertible Neural Networks for Medical Imaging

PatchNR: Learning from Small Data by Patch Normalizing Flow Regularization

Contact Info

Product

Resources

About