Cone-beam computed tomography (CBCT) for 3D orthopedic imaging produces excellent image quality of soft tissue and bone structures. Of particular interest is the ability for CBCT to acquire weight-bearing 3D images of the lower extremities, enabling examination under natural loading conditions. For patients experiencing pain, acquisition time should be minimized to reduce potential for motion. Toward this aim, we propose sparse-view acquisition sampling coupled with deep learning methods designed to reduce associated artifacts.

This presentation is part of Minisymposium “MS58 - Instruments and techniques for biomedical research (2 parts)”
organized by: Alberto Leardini (Laboratory of Movement Analysis and Functional-Clinical Evaluation of Prosthesis, Istituto Ortopedico Rizzoli, Bologna) .

Authors:

William Sehnert (Carestream Health Inc., Rochester)

Jiebo Luo (Department of Computer Science, University of Rochester)

Yuan Lin (Carestream Health, Inc)

Haofu Liao (University of Rochester)

Weijian Li (University of Rochester)

Zhimin Huo (Carestream Health, Inc)

David Foos (Carestream health, Inc)

Keywords:

artifact reduction, computed tomography, cone-beam ct, deep learning, image enhancement, image reconstruction, sparse view