Deep-learning based joint estimation of dual-tracer PET image activity maps and clustering of time activity curves

Yiming Wan; Huihui Ye; Huafeng Liu

doi:10.1117/12.2580873

15 February 2021 Deep-learning based joint estimation of dual-tracer PET image activity maps and clustering of time activity curves

Yiming Wan, Huihui Ye, Huafeng Liu

Author Affiliations +

Proceedings Volume 11595, Medical Imaging 2021: Physics of Medical Imaging; 115953T (2021) https://doi.org/10.1117/12.2580873
Event: SPIE Medical Imaging, 2021, Online Only

Conference Poster

Abstract

Dual-tracer positron emission tomography (PET) has emerged as a promising nuclear medical imaging technology. It reflects more information about biological functions than only using single tracer, which is of great value for disease research, clinical diagnosis and treatment. Furthermore, image reconstruction and segmentation often are treated as two sequential problems. This paper integrates these two problems into a coherent framework. In recent years, deep learning has attracted growing interests in medical image processing due to its powerful feature extraction ability. The reconstruction of simultaneous dual-tracer activity map based on neural network has shown an advantage of not relying on prior information and interval injection that conventional methods always need. In this paper, we propose a joint deep neural network for reconstruction of two individual tracer and segmentation by clustering of time activity curves. In the part of reconstruction, a classic Generative Adversarial Networks (GAN) called pix2pix is introduced as the basic framework. A 3D U-Net is used as the generator realize reconstruction. Then the time activity curves (TACs) in the reconstruction results are extracted and used for clustering to achieve segmentation. The loss of segmentation is used to guide the training of reconstruction network in return. The performance of segmentation network exceeds that of the previous joint method and the performance of reconstruction network is further improved.

Conference Presentation

Citation Download Citation

Yiming Wan, Huihui Ye, and Huafeng Liu "Deep-learning based joint estimation of dual-tracer PET image activity maps and clustering of time activity curves", Proc. SPIE 11595, Medical Imaging 2021: Physics of Medical Imaging, 115953T (15 February 2021); https://doi.org/10.1117/12.2580873

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