Feature Loss After Denoising of SPECT Projection Data using a U-Net

Reymann M, Massanes F, Ritt P, Cachovan M, Kuwert T, Vija AH, Maier A (2020)

Publication Language: English

Publication Type: Conference contribution, Abstract of a poster

Publication year: 2020

Conference Proceedings Title: Feature Loss After Denoising of SPECT Projection Data using a U-Net

Event location: Boston, Massachusetts

DOI: 10.1109/nss/mic42677.2020.9508041

Abstract

It has been long known that denoising of SPECT projection data from multi channel collimators harbors the danger that features are  filtered away as distinction between noise and signal in the 2D projection image does not consider the tomographic information (see Vija, Yahil et al.). We investigate if an approach using Neural Networks can overcome the past issues, by training a U-Net on Monte Carlo Simulations of the XCAT phantom. Although error metrics indicate improvement of signal statistics, we observe a loss of image features after the reconstruction. We show that when using OSEM reconstruction without denoising in projection domain, these features can be preserved.

Authors with CRIS profile

How to cite

APA:

Reymann, M., Massanes, F., Ritt, P., Cachovan, M., Kuwert, T., Vija, A.H., & Maier, A. (2020, November). Feature Loss After Denoising of SPECT Projection Data using a U-Net. Poster presentation at 2020 IEEE Nuclear Science Symposium and Medical Imaging Conference, Boston, Massachusetts, US.

MLA:

Reymann, Maximilian, et al. "Feature Loss After Denoising of SPECT Projection Data using a U-Net." Presented at 2020 IEEE Nuclear Science Symposium and Medical Imaging Conference, Boston, Massachusetts Ed. IEEE, 2020.

BibTeX: Download