ENHANCING HARDI RECONSTRUCTION FROM UNDERSAMPLED DATA VIA MULTI-CONTEXT AND FEATURE INTER-DEPENDENCY GAN

被引：7

作者：

Jha, Ranjeet Ranjan ^{[1
]}

Gupta, Hritik ^{[1
]}

Pathak, Sudhir K. ^{[2
]}

Schneider, Walter ^{[2
]}

Kumar, B. V. Rathish ^{[3
]}

Bhavsar, Arnav ^{[1
]}

Nigam, Aditya ^{[1
]}

机构：

[1] Indian Inst Technol Mandi, Sch Comp & Elect Engn SCEE, Mandi, Himachal Prades, India

[2] Univ Pittsburgh, Learning Res & Dev Ctr, Pittsburgh, PA 15260 USA

[3] Indian Inst Technol Kanpur, Dept Math & Stat, Kanpur, Uttar Pradesh, India

来源：

2021 IEEE 18TH INTERNATIONAL SYMPOSIUM ON BIOMEDICAL IMAGING (ISBI) | 2021年

关键词：

Diffusion MRI; Single/Multi Shell HARDI; Deep Learning; Generative Adversarial Network;

D O I：

10.1109/ISBI48211.2021.9434162

中图分类号：

R318 [生物医学工程];

学科分类号：

0831 ;

摘要：

In addition to the more traditional diffusion tensor imaging (DTI), over time, reconstruction techniques like HARDI have been proposed, which have a comparatively higher scanning time due to increased measurements, but are significantly better in the estimation of fiber structures. In order to make HARDI-based analysis faster, we propose an approach to reconstruct more MARDI volumes in q-space. The proposed GAN-based architecture leverages several modules, including a multi-context module, feature inter-dependencies module along-with numerous losses such as L1, adversarial, and total variation loss, to learn the transformation. The method is backed by some encouraging quantitative and visual results.

引用

页码：1103 / 1106

页数：4