English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The subject of "fairness" in artificial intelligence (AI) refers to assessingAI algorithms for potential bias based on demographic characteristics such asrace and gender, and the development of algorithms to address this bias. Mostapplications to date have been in computer vision, although some work inhealthcare has started to emerge. The use of deep learning (DL) in cardiac MRsegmentation has led to impressive results in recent years, and such techniquesare starting to be translated into clinical practice. However, no work has yetinvestigated the fairness of such models. In this work, we perform such ananalysis for racial/gender groups, focusing on the problem of training dataimbalance, using a nnU-Net model trained and evaluated on cine short axiscardiac MR data from the UK Biobank dataset, consisting of 5,903 subjects from6 different racial groups. We find statistically significant differences inDice performance between different racial groups. To reduce the racial bias, weinvestigated three strategies: (1) stratified batch sampling, in which batchsampling is stratified to ensure balance between racial groups; (2) fairmeta-learning for segmentation, in which a DL classifier is trained to classifyrace and jointly optimized with the segmentation model; and (3) protected groupmodels, in which a different segmentation model is trained for each racialgroup. We also compared the results to the scenario where we have a perfectlybalanced database. To assess fairness we used the standard deviation (SD) andskewed error ratio (SER) of the average Dice values. Our results demonstratethat the racial bias results from the use of imbalanced training data, and thatall proposed bias mitigation strategies improved fairness, with the best SD andSER resulting from the use of protected group models.

Fairness in Cardiac MR Image Analysis: An Investigation of Bias Due to Data Imbalance in Deep Learning Based Segmentation