Enhancing signal detectability in learning-based CT reconstruction with a model observer inspired loss function

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Lantz, Megan, Sidky, Emil Y., Reiser, Ingrid S., Pan, Xiaochuan, Ongie, Gregory
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916126656561152
author Lantz, Megan
Sidky, Emil Y.
Reiser, Ingrid S.
Pan, Xiaochuan
Ongie, Gregory
author_facet Lantz, Megan
Sidky, Emil Y.
Reiser, Ingrid S.
Pan, Xiaochuan
Ongie, Gregory
contents 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 pixel-wise 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.
format Preprint
id arxiv_https___arxiv_org_abs_2402_10010
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Enhancing signal detectability in learning-based CT reconstruction with a model observer inspired loss function
Lantz, Megan
Sidky, Emil Y.
Reiser, Ingrid S.
Pan, Xiaochuan
Ongie, Gregory
Medical Physics
Computer Vision and Pattern Recognition
Image and Video Processing
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 pixel-wise 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.
title Enhancing signal detectability in learning-based CT reconstruction with a model observer inspired loss function
topic Medical Physics
Computer Vision and Pattern Recognition
Image and Video Processing
url https://arxiv.org/abs/2402.10010