Learning to recover sharp detail from simulated low-dose ct studies

Patrick Cole, Ayis Pyrros, Oluwasanmi Koyejo

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Radiology exams require exposing a patient to a variable dosage of radiation. Importantly, the amount of radiation used during the exam directly corresponds to the level of noise in the resulting image, and increased amounts of radiation can pose health risks to patients. This results in a tradeoff, as radiologists need a high-quality image to make a diagnosis. In this work, we propose a method to recover image fidelity given a noisy, or low-dose, sample. Using a two-part criterion that consists of a pixel-wise loss and an adversarial loss, we are able to recover the structure and fine detail of the normal-dose sample. To evaluate the denoising method, we implement simulations of realistic low-dose noise for a computed tomography exam, which may be of independent interest. Quantitative and qualitative results highlight the performance of our approach as compared to existing baselines.

Original languageEnglish (US)
Title of host publication2021 IEEE 18th International Symposium on Biomedical Imaging, ISBI 2021
PublisherIEEE Computer Society
Pages748-752
Number of pages5
ISBN (Electronic)9781665412469
DOIs
StatePublished - Apr 13 2021
Externally publishedYes
Event18th IEEE International Symposium on Biomedical Imaging, ISBI 2021 - Nice, France
Duration: Apr 13 2021Apr 16 2021

Publication series

NameProceedings - International Symposium on Biomedical Imaging
Volume2021-April
ISSN (Print)1945-7928
ISSN (Electronic)1945-8452

Conference

Conference18th IEEE International Symposium on Biomedical Imaging, ISBI 2021
Country/TerritoryFrance
CityNice
Period4/13/214/16/21

Keywords

  • Computed tomography
  • Computer vision
  • Deep learning
  • Image denoising
  • Machine learning
  • Radiology

ASJC Scopus subject areas

  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging

Fingerprint

Dive into the research topics of 'Learning to recover sharp detail from simulated low-dose ct studies'. Together they form a unique fingerprint.

Cite this