Share Email Print
cover

Proceedings Paper • new

Reduction of metal artifacts in x-ray CT images using a convolutional neural network
Author(s): Yanbo Zhang; Ying Chu; Hengyong Yu
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Patients usually contain various metallic implants (e.g. dental fillings, prostheses), causing severe artifacts in the x-ray CT images. Although a large number of metal artifact reduction (MAR) methods have been proposed in the past four decades, MAR is still one of the major problems in clinical x-ray CT. In this work, we develop a convolutional neural network (CNN) based MAR framework, which combines the information from the original and corrected images to suppress artifacts. Before the MAR, we generate a group of data and train a CNN. First, we numerically simulate various metal artifacts cases and build a dataset, which includes metal-free images (used as references), metal-inserted images and various MAR methods corrected images. Then, ten thousands patches are extracted from the databased to train the metal artifact reduction CNN. In the MAR stage, the original image and two corrected images are stacked as a three-channel input image for CNN, and a CNN image is generated with less artifacts. The water equivalent regions in the CNN image are set to a uniform value to yield a CNN prior, whose forward projections are used to replace the metal affected projections, followed by the FBP reconstruction. Experimental results demonstrate the superior metal artifact reduction capability of the proposed method to its competitors.

Paper Details

Date Published: 25 September 2017
PDF: 11 pages
Proc. SPIE 10391, Developments in X-Ray Tomography XI, 103910V (25 September 2017); doi: 10.1117/12.2275592
Show Author Affiliations
Yanbo Zhang, Univ. of Massachusetts Lowell (United States)
Ying Chu, Univ. of Massachusetts Lowell (United States)
Shenzhen Univ. (China)
Hengyong Yu, Univ. of Massachusetts Lowell (United States)


Published in SPIE Proceedings Vol. 10391:
Developments in X-Ray Tomography XI
Bert Müller; Ge Wang, Editor(s)

© SPIE. Terms of Use
Back to Top