2021
DOI: 10.3390/jimaging7120248
|View full text |Cite
|
Sign up to set email alerts
|

Learned Primal Dual Reconstruction for PET

Abstract: We have adapted, implemented and trained the Learned Primal Dual algorithm suggested by Adler and Öktem and evaluated its performance in reconstructing projection data from our PET scanner. Learned Primal Dual reconstructions are compared to Maximum Likelihood Expectation Maximisation (MLEM) reconstructions. Different strategies for training are also compared. Whenever the noise level of the data to reconstruct is sufficiently represented in the training set, the Learned Primal Dual algorithm performs well on … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1

Citation Types

0
2
0

Year Published

2023
2023
2024
2024

Publication Types

Select...
3
3

Relationship

0
6

Authors

Journals

citations
Cited by 6 publications
(2 citation statements)
references
References 14 publications
0
2
0
Order By: Relevance
“… LPD 58 when applied to example PET phantom data. 59 Figure adapted from data courtesy Massimiliano Colarieti-Tosti. The overall processing pipeline effectively has two rails: the top rail conducts sequential processing of the sinogram data and the lower rail conducts sequential processing of the image.…”
Section: Ai Methods: From Direct To Iterativementioning
confidence: 99%
“… LPD 58 when applied to example PET phantom data. 59 Figure adapted from data courtesy Massimiliano Colarieti-Tosti. The overall processing pipeline effectively has two rails: the top rail conducts sequential processing of the sinogram data and the lower rail conducts sequential processing of the image.…”
Section: Ai Methods: From Direct To Iterativementioning
confidence: 99%
“…The LPD was adapted by using the OSEM reconstruction as initialisation of primal channels, and by including an affine forward operator with sample-specific multiplicative and additive factors. These sample-specific factors are not included in previous LPD implementations for PET [4]. To reduce computational time a total of 3 unrolled iterations were used.…”
Section: A Deep Learning Methodsmentioning
confidence: 99%