eduzhai > Applied Sciences > Engineering >

Grading Loss A Fracture Grade-based Metric Loss for Vertebral Fracture Detection

  • king
  • (0) Download
  • 20210506
  • Save

... pages left unread,continue reading

Document pages: 11 pages

Abstract: Osteoporotic vertebral fractures have a severe impact on patients overallwell-being but are severely under-diagnosed. These fractures present themselvesat various levels of severity measured using the Genant s grading scale.Insufficient annotated datasets, severe data-imbalance, and minor difference inappearances between fractured and healthy vertebrae make naive classificationapproaches result in poor discriminatory performance. Addressing this, wepropose a representation learning-inspired approach for automated vertebralfracture detection, aimed at learning latent representations efficient forfracture detection. Building on state-of-art metric losses, we present a novelGrading Loss for learning representations that respect Genant s fracturegrading scheme. On a publicly available spine dataset, the proposed lossfunction achieves a fracture detection F1 score of 81.5 , a 10 increase over anaive classification baseline.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×