Medical applications of example-based super-resolution

Principal Investigator

Name
Ramin Zabih

Degrees
Ph.D.

Institution
Joan & Sanford I. Weill Medical College of Cornell University

Position Title
Associate Professor of Computer Science in Radiology

Email
rdz@med.cornell.edu

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-279

Initial CDAS Request Approval
Jan 31, 2017

Title
Medical applications of example-based super-resolution

Summary
Example-based super-resolution (EBSR) reconstructs a high-resolution image from a low-resolution image, given a training set of high-resolution images. In this project I plan to implement some applications of EBSR to medical imaging. A particular interesting application, which I call "x-ray voxelization", approximates the result of a CT scan from an x-ray image. This is useful for reducing radiation dose received by patients and can dramatically improve dose-controlled precision.

Aims

In the CT arm of NLST, we plan to treat the localizer scout image (taken by the CT technologist prior to the main scan) like a chest radiograph. We will then use the axial images from the helical CT scan to train the x-ray voxelization algorithm, and compare with the results of existing x-ray voxelization algorithms.

Collaborators

Akshay Bhat, Cornell University
Chen Wang, Cornell University
Charles Herrmann, Cornell University