Skip to content

webirooni

Half Mag / Half Zine

  • News
  • Business
  • Technology
  • Lifestyle
  • Sport
  • Shop
  • Privacy Policy
  • T & C
  • Contact Us
  • Toggle search form

Deep learning for extremity radiographs confounded by labels

Posted on November 22, 2021November 20, 2021 By webiroonicom

According to an open-access Editor’s Choice article in ARRS’ American Journal of Roentgenology (AJR), convolutional neural networks (CNN) trained to identify abnormalities on upper extremity radiographs are susceptible to a ubiquitous confounding image feature that could limit their clinical utility: radiograph labels.

“We recommend that such potential image confounders be collected when possible during dataset curation, and that covering these labels be considered during CNN training,” wrote corresponding author Paul H. Yi from the University of Maryland’s Medical Intelligent Imaging Center in Baltimore.

Yi and team’s retrospective study evaluated 40,561 upper extremity musculoskeletal radiographs from Stanford’s MURA dataset that were used to train three DenseNet-121 CNN classifiers. Three inputs were used to distinguish normal from abnormal radiographs: original images with both anatomy and labels; images with laterality and/or technologist labels subsequently covered by a black box; images where anatomy had been removed and only labels remained.

For the original radiographs, AUC was 0.844, frequently emphasizing laterality and/or technologist labels for decision-making. Covering these labels increased AUC to 0.857 (p=.02) and redirected CNN attention from the labels to the bones. Using labels alone, AUC was 0.638, indicating that radiograph labels are associated with abnormal examinations.

“While we can infer that labels are associated with normal versus abnormal disease categories,” the authors of this AJR article added, “we cannot determine the specific aspect of the labels that resulted in their being confounding factors.”

Lifestyle

Post navigation

Previous Post: Airbus shaves 20-year demand forecast, sees faster replacements
Next Post: Thursday Night Football: Big defensive night for Dolphins in 22-10 win over Ravens

More Related Articles

Solo seniors with social support are less likely to need nursing home care Lifestyle
Common blood pressure drug does not slow down the progression of more advanced Alzheimer’s, study finds Lifestyle
Lower risk of death for Black Veterans with heart failure and pneumonia in VA system Lifestyle
A key brain region responds to faces similarly in infants and adults Lifestyle
Evidence dyslexia affects children’s visual processing beyond just reading Lifestyle
Researchers target a mouse’s own cells, rather than using antibiotics, to treat pneumonia Lifestyle

Recent Posts

  • Mike White: If you ask me, I should have been a first overall pick
  • Lower risk of death for Black Veterans with heart failure and pneumonia in VA system
  • Jetmakers push freighters as industry gathers in Dubai
  • Discord CEO says the company won’t add a crypto wallet for now
  • Penguin found 1,800 miles from home on New Zealand beach

Categories

  • Business
  • Lifestyle
  • News
  • Sport
  • Technology

Copyright © 2023 webirooni.

Powered by PressBook Blog WordPress theme