PET image resolution uniformity improvements using deep learning

Jing Lin; Yingying Li; Huihui Ye; Huafeng Liu

doi:10.1117/12.2580637

15 February 2021 PET image resolution uniformity improvements using deep learning

Jing Lin, Yingying Li, Huihui Ye, Huafeng Liu

Proceedings Volume 11595, Medical Imaging 2021: Physics of Medical Imaging; 115954A (2021) https://doi.org/10.1117/12.2580637
Event: SPIE Medical Imaging, 2021, Online Only

Conference Poster

Abstract

The position-coding errors of the lines of response (LORs) caused by the depth of interaction (DOI) effects, especially the radial DOI effect, leads to the nonuniform resolution of the determined field of view (FOV) in positron emission tomography (PET), so that severe “tailing” will appear in the reconstructed image. At present, the most commonly used approach to solve the radial DOI problem is the hardware method, which mainly includes designing new detectors by using different principles. However, those approaches rely one more complex detector structures and signal processing methods. Inspired by systems that employ machine learning at the hardware level, a new deep learning based approach dedicated for improving the uniformity of radial resolution in PET imaging is proposed, whose basic idea is that images with high resolution at the center of FOV are fed as labels into ISTA-Net to train the sinograms, which are from the same objects at the edge of the FOV. The network makes use of the nonuniformity itself without requiring new detector or obtaining any additional information, such as DOI information from specified detectors or point spread function (PSF) information. Qualitative assessment and quantitative analysis based on real rats dataset and Monte Carlo simulations where four-layered DOI coding system is designed to obtain DOI information demonstrated that the proposed reconstruction method greatly improves the resolution of the image at the edge of FOV and enhances the uniformity of radial resolution in detection space.

Conference Presentation

Citation Download Citation

Jing Lin, Yingying Li, Huihui Ye, and Huafeng Liu "PET image resolution uniformity improvements using deep learning", Proc. SPIE 11595, Medical Imaging 2021: Physics of Medical Imaging, 115954A (15 February 2021); https://doi.org/10.1117/12.2580637

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available