University of Illinois Urbana-Champaign

Increasing trust through the design of algorithm-based lesion segmentation support systems

Gryska, Emilia; Cerna, Katerina; Heckemann, Rolf A.

Permalink

https://hdl.handle.net/2142/106560

Description

Title
Increasing trust through the design of algorithm-based lesion segmentation support systems
Author(s)
  • Gryska, Emilia
  • Cerna, Katerina
  • Heckemann, Rolf A.
Issue Date
2020-03-23
Keyword(s)
  • Algorithm-based support systems
  • Brain lesion segmentation
  • Trust
  • Design for clinical practice
Abstract
The adoption rate of algorithm-based lesion segmentation support systems in clinical practice is very low. This is partly due to low trust levels radiologists have in such systems. To increase the trust, the design and validation of the support tools must comply with the needs and expectations of radiologists. We interviewed four clinicians who work with brain images on a daily basis to understand the needs, current methods and practices of image interpretation, and their opinion of automatic brain lesion segmentation tools. In the interviews, we identified the necessity to state the error of the automated decision support tool and its clinical relevance in a given context.
Publisher
iSchools
Series/Report Name or Number
iConference 2020 Proceedings
Type of Resource
  • text
  • image
Language
eng
Permalink
http://hdl.handle.net/2142/106560
Copyright and License Information
Copyright 2020 Emilia Gryska, Katerina Cerna, and Rolf A. Heckemann

Owning Collections

iConference 2020 Poster Descriptions PRIMARY