TY - JOUR
T1 - Quasi-supervised learning for super-resolution PET
AU - Yang, Guangtong
AU - Li, Chen
AU - Yao, Yudong
AU - Wang, Ge
AU - Teng, Yueyang
N1 - Publisher Copyright:
© 2024 Elsevier Ltd
PY - 2024/4
Y1 - 2024/4
N2 - Low resolution of positron emission tomography (PET) limits its diagnostic performance. Deep learning has been successfully applied to achieve super-resolution PET. However, commonly used supervised learning methods in this context require many pairs of low- and high-resolution (LR and HR) PET images. Although unsupervised learning utilizes unpaired images, the results are not as good as that obtained with supervised deep learning. In this paper, we propose a quasi-supervised learning method, which is a new type of weakly-supervised learning methods, to recover HR PET images from LR counterparts by leveraging similarity between unpaired LR and HR image patches. Specifically, LR image patches are taken from a patient as inputs, while the most similar HR patches from other patients are found as labels. The similarity between the matched HR and LR patches serves as a prior for network construction. Our proposed method can be implemented by designing a new network or modifying an existing network. As an example in this study, we have modified the cycle-consistent generative adversarial network (CycleGAN) for super-resolution PET. Our numerical and experimental results qualitatively and quantitatively show the merits of our method relative to the state-of-the-art methods. The code is publicly available at https://github.com/PigYang-ops/CycleGAN-QSDL.
AB - Low resolution of positron emission tomography (PET) limits its diagnostic performance. Deep learning has been successfully applied to achieve super-resolution PET. However, commonly used supervised learning methods in this context require many pairs of low- and high-resolution (LR and HR) PET images. Although unsupervised learning utilizes unpaired images, the results are not as good as that obtained with supervised deep learning. In this paper, we propose a quasi-supervised learning method, which is a new type of weakly-supervised learning methods, to recover HR PET images from LR counterparts by leveraging similarity between unpaired LR and HR image patches. Specifically, LR image patches are taken from a patient as inputs, while the most similar HR patches from other patients are found as labels. The similarity between the matched HR and LR patches serves as a prior for network construction. Our proposed method can be implemented by designing a new network or modifying an existing network. As an example in this study, we have modified the cycle-consistent generative adversarial network (CycleGAN) for super-resolution PET. Our numerical and experimental results qualitatively and quantitatively show the merits of our method relative to the state-of-the-art methods. The code is publicly available at https://github.com/PigYang-ops/CycleGAN-QSDL.
KW - Positron emission tomography (PET)
KW - Super-resolution
KW - Unpaired data
KW - Weakly-supervised learning
UR - http://www.scopus.com/inward/record.url?scp=85184768555&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85184768555&partnerID=8YFLogxK
U2 - 10.1016/j.compmedimag.2024.102351
DO - 10.1016/j.compmedimag.2024.102351
M3 - Article
C2 - 38335784
AN - SCOPUS:85184768555
SN - 0895-6111
VL - 113
JO - Computerized Medical Imaging and Graphics
JF - Computerized Medical Imaging and Graphics
M1 - 102351
ER -