Medical Image Synthesis with Context-Aware Generative Adversarial Networks

@article{Nie2017MedicalIS,
  title={Medical Image Synthesis with Context-Aware Generative Adversarial Networks},
  author={Dong Nie and Roger Trullo and Caroline Petitjean and Su Ruan and Dinggang Shen},
  journal={Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention},
  year={2017},
  volume={10435},
  pages={
          417-425
        }
}
  • D. Nie, Roger Trullo, D. Shen
  • Published 16 December 2016
  • Computer Science
  • Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Intervention
Computed tomography (CT) is critical for various clinical applications, e.g., radiation treatment planning and also PET attenuation correction in MRI/PET scanner. [] Key Method Specifically, we train a fully convolutional network (FCN) to generate CT given the MR image. To better model the nonlinear mapping from MRI to CT and produce more realistic images, we propose to use the adversarial training strategy to train the FCN.
GAN for synthesizing CT from T2-weighted MRI data towards MR-guided radiation treatment
TLDR
A conditional generative adversarial network with multiple residual blocks to estimate sCT from T2-weighted MRI scans using 367 paired brain MR-CT images dataset is proposed and qualitatively evaluated by visual comparisons of generated sCT to original CT of respective MRI input.
Medical Image Synthesis with Deep Convolutional Adversarial Networks
TLDR
This paper trains a fully convolutional network to generate a target image given a source image and proposes to use the adversarial learning strategy to better model the FCN, designed to incorporate an image-gradient-difference-based loss function to avoid generating blurry target images.
CT-Based MR Synthesis Using Adversarial Cycle-Consistent Networks with Paired Data Learning
  • Cheng-Bin Jin, Hakil Kim, X. Cui
  • Medicine
    2018 11th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI)
  • 2018
TLDR
A synthetic approach that translates a CT image into an MR image using adversarial cycle-consistent networks with paired data learning that is accurate and robust for predicting MR images from CT images and also outperforms two state-of-the-art methods.
A feature invariant generative adversarial network for head and neck MRI/CT image synthesis
TLDR
A novel deep image synthesis model that generates in an unsupervised manner CT images from diagnostic MRI for radiotherapy planning by learning a new invariant representation to generate synthetic CT images based on high frequency and appearance patterns.
MRI-only based synthetic CT generation using dense cycle consistent generative adversarial networks.
TLDR
A novel learning-based approach to generate CT images from routine MRIs based on dense cycle GAN model to effectively capture the relationship between the CT and MRIs and offers strong potential for supporting near real-time MRI-only treatment planning in the brain and pelvis.
Evaluation of CT Image Synthesis Methods:From Atlas-based Registration to Deep Learning
TLDR
It is demonstrated that image synthesis methods allow for more accurate bone segmentation using the synthetic CT imaging than would otherwise be possible by segmenting the bone in the MRI directly.
Structurally aware bidirectional unpaired image to image translation between CT and MR
TLDR
Two different variations of Generative Adversarial Networks are implemented exploiting the cycling consistency and structural similarity between both CT and MR image modalities on a pelvis dataset, thus facilitating a bidirectional exchange of content and style between these imagemodalities.
Frequency-Supervised MR-to-CT Image Synthesis
TLDR
A frequency decomposition layer that learns to decompose predicted CT outputs into low and highfrequency components, and a refinement module to improve high-frequency reconstruction through high- frequencies adversarial learning are proposed.
MRI-based synthetic CT generation using deep convolutional neural network
TLDR
A novel deep learning-based method to generate sCT with a high accuracy that makes the sCT comparable to that of the planning CT, which could be a useful tool for MRI-based radiation treatment planning and attenuation correction in a PET/MRI scanner.
Capturing variabilities from Computed Tomography images with Generative Adversarial Networks (GANs)
TLDR
This work proposes a novel unsupervised data-driven approach for data augmentation that can generate 2D Computed Tomography (CT) images using a simple GAN and shows that the proposed solution using GANs is able to capture some of the global and local CT variabilities.
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 26 REFERENCES
Estimating CT Image From MRI Data Using Structured Random Forest and Auto-Context Model
TLDR
Experimental results show that the learning-based method proposed can accurately predict CT images in various scenarios, even for the images undergoing large shape variation, and also outperforms two state-of-the-art methods.
Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network
TLDR
SRGAN, a generative adversarial network (GAN) for image super-resolution (SR), is presented, to its knowledge, the first framework capable of inferring photo-realistic natural images for 4x upscaling factors and a perceptual loss function which consists of an adversarial loss and a content loss.
Improving magnetic resonance resolution with supervised learning
TLDR
An automated supervised learning algorithm is presented that is similar to the brain hallucination work of Rousseau, taking advantage of new developments in regression based image reconstruction, demonstrating the improvement over state-of-the-art super-resolution techniques.
Deep Learning in Medical Image Analysis.
TLDR
This review covers computer-assisted analysis of images in the field of medical imaging and introduces the fundamentals of deep learning methods and their successes in image registration, detection of anatomical and cellular structures, tissue segmentation, computer-aided disease diagnosis and prognosis, and so on.
LINKS: Learning-Based Multi-source IntegratioN FrameworK for Segmentation of Infant Brain Images
TLDR
Experimental results show that the proposed method achieves better performance than other state-of-the-art automated segmentation methods, with significantly reduction of running time from hours to 5 minutes.
Attenuation correction for a combined 3D PET/CT scanner.
TLDR
The proof of principle of CT-based attenuation correction of 3D positron emission tomography (PET) data is demonstrated by using scans of bone and soft tissue equivalent phantoms and scans of humans to conclude that using CT information is a feasible way to obtain attenuation Correction factors for 3D PET.
Deep multi-scale video prediction beyond mean square error
TLDR
This work trains a convolutional network to generate future frames given an input sequence and proposes three different and complementary feature learning strategies: a multi-scale architecture, an adversarial training method, and an image gradient difference loss function.
Fully convolutional networks for multi-modality isointense infant brain image segmentation
TLDR
This paper proposes to use fully convolutional networks (FCNs) for the segmentation of isointense phase brain MR images, and shows that the proposed model significantly outperformed previous methods in terms of accuracy and indicated a better way of integrating multi-modality images, which leads to performance improvement.
Deep Learning Based Imaging Data Completion for Improved Brain Disease Diagnosis
TLDR
This work proposed a deep learning based framework for estimating multi-modality imaging data in the form of convolutional neural networks, where the input and output are two volumetric modalities.
Image Super-Resolution Using Deep Convolutional Networks
We propose a deep learning method for single image super-resolution (SR). Our method directly learns an end-to-end mapping between the low/high-resolution images. The mapping is represented as a deep
...
1
2
3
...