MDPET: A Unified Motion Correction and Denoising Adversarial Network for Low-Dose Gated PET

Zhou, Bo; Tsai, Yi‐Fang; Chen, X; Duncan, James S.; Liu, Chi

doi:10.1109/tmi.2021.3076191

Cited by 36 publications

(10 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the number of institutions could be more than three with more diverse low-dose settings. For example, Kaplan and Zhu [5], Ladefoged et al [50], and Zhou et al [51] used a 10%-count low-dose setting. Wang et al [3] and Whiteley et al [52] deployed a 25%-count low-dose setting, while Ouyang et al [9] considered a 1%count ultralow-dose setting.…”

Section: Discussionmentioning

confidence: 99%

Federated Transfer Learning for Low-Dose PET Denoising: A Pilot Study With Simulated Heterogeneous Data

Zhou

Miao

Mirian

et al. 2023

IEEE Trans. Radiat. Plasma Med. Sci.

Self Cite

View full text Add to dashboard Cite

Positron emission tomography (PET) with a reduced injection dose, i.e., low-dose PET, is an efficient way to reduce radiation dose. However, low-dose PET reconstruction suffers from a low signal-to-noise ratio (SNR), affecting diagnosis and other PET-related applications. Recently, deep learning-based PET denoising methods have demonstrated superior performance in generating high-quality reconstruction. However, these methods require a large amount of representative data for training, which can be difficult to collect and share due to medical data privacy regulations. Moreover, low-dose PET data at different institutions may use different low-dose protocols, leading to nonidentical data distribution. While previous federated learning (FL) algorithms enable multiinstitution collaborative training without the need of aggregating local data, it is challenging for previous methods to address the large domain shift caused by different low-dose PET settings, and the application of FL to PET is still underexplored. In this work, we propose a federated transfer learning (FTL) framework for low-dose PET denoising using heterogeneous low-dose data. Our experimental results on simulated multiinstitutional data demonstrate that our method can efficiently utilize heterogeneous low-dose data without compromising data privacy for achieving superior low-dose PET

show abstract

Section: Discussionmentioning

confidence: 99%

Federated Transfer Learning for Low-Dose PET Denoising: A Pilot Study With Simulated Heterogeneous Data

Zhou

Miao

Mirian

et al. 2023

IEEE Trans. Radiat. Plasma Med. Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Zhou et al. [25] proposed a unified motion correction and denoising adversarial network for generating motion‐compensated low noise images from low‐dose gated PET data. In recent years, Kupyn et al.…”

Section: Related Workmentioning

confidence: 99%

“…Experimental results showed that their model can generate images conditioned on class labels. Zhou et al [25] proposed a unified motion correction and denoising adversarial network for generating motion-compensated low noise images from low-dose gated PET data. In recent years, Kupyn et al proposed a deblurring Gan model named Deblur-Gan [26], which trains a recovery image generator using CNN technology and a discriminant network patchGan, and the loss function is composed of the counter loss and content loss.…”

Section: Related Workmentioning

confidence: 99%

Blood‐contaminated endoscopic image restoration based on residual VQ‐VAE with cascaded structure

Lin

Pan

Yuan

et al. 2022

IET Image Processing

View full text Add to dashboard Cite

This research studies an image reconstruction technology that can help doctors observe endoscopic images obscured by blood in transurethral procedures. A new network structure is proposed, which applies encoder–decoder deep neural network to achieve the purpose of denoising and uses residual learning and cascade learning methods to relieve the problem of insufficient clarity and lack of detail in reconstruction image when restoring endoscopic images suffered from blood occlusion. The residual mechanism is used to estimate noise and compensate the images, instead of reconstructing overall image by an encoder and decoder deep network. It can potentially make the details of the image to be preserved and improve the resolution of the reconstructed image. This denoising model is embedded into a cascade network framework to further improve the quality of the restored image through the idea of repeated learning multiple times. On a customized endoscopic image sample set with real blood noise, for verifying the denoising ability, we designed two approaches of data augmentation to generate noised images obscured by different types and degree blood noises. In experiments, the new model has been compared to other approaches and the certain improved denoising effects can be observed.

show abstract

“…al. 15 , also proposed a unified motion correction and denoising deep framework for low-dose gated PET. In our work, we implemented a direct motion correction framework in imagedomain based on U-Net architecture without the need to access PET raw data.…”

Section: Introductionmentioning

confidence: 99%

Direct respiratory motion correction of whole-body PET images using a deep learning framework incorporating spatial information

Miao

Tsai

et al. 2023

Medical Imaging 2023: Physics of Medical Imaging

View full text Add to dashboard Cite

Breathing can cause blurring and artifacts in PET images, especially in the body trunk region. The blurring and artifacts can negatively impact cancer detection and response to therapy assessment. Many motion detection techniques, such as using external motion sensors or data driving methods, have been used to facilitate respiratory motion correction for PET. These methods require sophisticated gating or motion compensated image reconstruction, which are time consuming. In our work, we propose a deep learning framework based on U-Net to directly perform respiratory motion correction for whole-body PET in the image domain. Our framework was trained with the patches of the PET images without motion correction as input and the motion-corrected images using the data-driving gating (DDG) method as label. The framework also incorporated spatial information of voxels by adding additional layer of voxels’ vertical locations to the input. To evaluate our framework, we conducted 5-fold cross validations to generate the motion-corrected images for 30 subjects and compared them with the ground truth images corrected by the DDG methods. Our framework could correct the PET images in regions affected by respiratory motion. The incorporation of voxels’ spatial information could further improve the performance of motion correction. There is a great potential of our framework to perform direct respiratory motion correction in the image domain in a convenient manner.

show abstract

MDPET: A Unified Motion Correction and Denoising Adversarial Network for Low-Dose Gated PET

Cited by 36 publications

References 42 publications

Federated Transfer Learning for Low-Dose PET Denoising: A Pilot Study With Simulated Heterogeneous Data

Federated Transfer Learning for Low-Dose PET Denoising: A Pilot Study With Simulated Heterogeneous Data

Blood‐contaminated endoscopic image restoration based on residual VQ‐VAE with cascaded structure

Direct respiratory motion correction of whole-body PET images using a deep learning framework incorporating spatial information

Contact Info

Product

Resources

About