Optimizing Model Observer Performance in Learning-Based CT Reconstruction

Authors

Gregory Ongie, Marquette UniversityFollow
Emil Y. Sidky, University of Chicago
Ingrid S. Reiser, University of Chicago
Xiaochuan Pan, University of Chicago

Document Type

Conference Proceeding

Publication Date

2022

Publisher

Society of Photo-optical Instrumentation Engineers (SPIE)

Source Publication

Proceedings of SPIE 12035: Medical Imaging

Source ISSN

9781510649460

Original Item ID

DOI: 10.1117/12.2613050

Abstract

Deep neural networks used for reconstructing sparse-view CT data are typically trained by minimizing a pixel- wise mean-squared error or similar loss function over a set of training images. However, networks trained with such losses are prone to wipe out small, low-contrast features that are critical for screening and diagnosis. To remedy this issue, we introduce a novel training loss inspired by the model observer framework to enhance the detectability of weak signals in the reconstructions. We evaluate our approach on the reconstruction of synthetic sparse-view breast CT data, and demonstrate an improvement in signal detectability with the proposed loss.

Comments

Published in Proceedings of SPIE 12035: Medical Imaging, 2022. DOI. © Society of Photo-optical Instrumentation Engineers (SPIE). Used with permission.

Recommended Citation

Ongie, Gregory; Sidky, Emil Y.; Reiser, Ingrid S.; and Pan, Xiaochuan, "Optimizing Model Observer Performance in Learning-Based CT Reconstruction" (2022). Mathematical and Statistical Science Faculty Research and Publications. 139.
https://epublications.marquette.edu/math_fac/139