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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Deep learning (DL) is gaining popularity as a parameter estimation method forquantitative MRI. A range of competing implementations have been proposed,relying on either supervised or self-supervised learning. Self-supervisedapproaches, sometimes referred to as unsupervised, have been loosely based onauto-encoders, whereas supervised methods have, to date, been trained ongroundtruth labels. These two learning paradigms have been shown to havedistinct strengths. Notably, self-supervised approaches have offered lower-biasparameter estimates than their supervised alternatives. This result iscounterintuitive - incorporating prior knowledge with supervised labels should,in theory, lead to improved accuracy. In this work, we show that this apparentlimitation of supervised approaches stems from the naive choice of groundtruthtraining labels. By training on labels which are deliberately not groundtruth,we show that the low-bias parameter estimation previously associated withself-supervised methods can be replicated - and improved on - within asupervised learning framework. This approach sets the stage for a single,unifying, deep learning parameter estimation framework, based on supervisedlearning, where trade-offs between bias and variance are made by carefuladjustment of training label.

Choice of training label matters: how to best use deep learning for quantitative MRI parameter estimation