Ground truth

Last revised by Dimitrios Toumpanakis on 18 Apr 2021

Citation, DOI, disclosures and article data

Citation:

Moore C, Toumpanakis D, Ground truth. Reference article, Radiopaedia.org (Accessed on 25 Apr 2024) https://doi.org/10.53347/rID-80984

DOI:

https://doi.org/10.53347/rID-80984

Permalink:

https://radiopaedia.org/articles/80984

rID:

80984

Article created:

11 Aug 2020, Candace Makeda Moore

Disclosures:

At the time the article was created Candace Makeda Moore had no recorded disclosures.

View Candace Makeda Moore's current disclosures

Last revised:

18 Apr 2021, Dimitrios Toumpanakis

Disclosures:

At the time the article was last revised Dimitrios Toumpanakis had no recorded disclosures.

View Dimitrios Toumpanakis's current disclosures

Revisions:

3 times, by 2 contributors - see full revision history and disclosures

Sections:

Artificial Intelligence

Ground truth is a term used in statistics and machine learning to refer to data assumed to be correct.

Regarding the development of machine learning algorithms in radiology, the ground truth for image labeling is sometimes based on pathology or lab results while, in other cases, on the expert opinion of experienced radiologists. In the computing informatics literature the term ground truth can be sometimes used to imply any data presumed to be correctly labeled, and thus it is possible to create synthetic ground truth ^3,4.

When developing a machine learning model using ground truth data, it is important to first explore/analyze "how true" the ground truth actually is, i.e. to compare the inter- and intra-observer variability in the data used as ground truth.