On ‘ground’ truth and why we should abandon the term

Woodhouse, Iain

doi:10.1117/1.jrs.15.041501

Cited by 8 publications

(5 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Second, given the differences in urban de nitions in these datasets, standard accuracy estimates may not be particularly helpful. There has been discussion about the term 'ground-truth' in the broader remote sensing community that is relevant here 74 . However, as a sanity check, we provide a basic accuracy estimates of the eight datasets used for representing present-day (2019/2020) urban land using the validation dataset created by the Dynamic World team 28 .…”

Section: Validationmentioning

confidence: 99%

Large disagreements in estimates of urban land across scales and their implications

Chakraborty,

Venter,

Demuzere

et al. 2024

Preprint

View full text Add to dashboard Cite

Improvements in high-resolution satellite remote sensing and computational advancements have sped up the development of global datasets that delineate urban land, crucial for understanding climate risks in our increasingly urbanizing world. Here, we analyze urban land cover patterns across spatiotemporal scales from several such current-generation products. While all the datasets show a rapidly urbanizing world, with global urban land nearly tripling between 1985 and 2015, there are substantial discrepancies in urban land area estimates among them influenced by scale, differing urban definitions, and methodologies. We discuss the implications of these discrepancies for several use cases, including for monitoring urban climate risks and for modeling urbanization-induced impacts on weather and climate from regional to global scales. Our results demonstrate the importance of choosing fit for purpose datasets for examining specific aspects of historical, present, and future urbanization with implications for sustainable development, resource allocation, and quantification of climate impacts.

show abstract

Section: Validationmentioning

confidence: 99%

Large disagreements in estimates of urban land across scales and their implications

Chakraborty,

Venter,

Demuzere

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…GTD for DF is described by Göbel et al. [8] as “data whose content is well known and understood by the community, i.e., the exact information or the actual digital traces to be discovered in a data set during the investigation are well documented.” While Woodhouse [1] suggests the origins of the term GTD are unclear, it is often linked with both military and cartography domains, and the field of DF has recently sought to adopt this term to describe data that has been methodically created and its content known and documented. The field of DF concerns itself with the examination of “digital traces” and those within it seek to determine the value and meaning of any that they consider are potentially relevant during the course of an investigation [15].…”

Section: Knowledge Discovery and Gtdmentioning

confidence: 99%

“…Ground truth data (GTD) is data that represent the truth for a given state or scenario [1]. This type of data is gathered from direct observations rather than generated from inferences [2] and can help us to understand how something works and the impact and characteristics of any given event.…”

Section: Introductionmentioning

confidence: 99%

A template for creating and sharing ground truth data in digital forensics

Horsman

2024

Journal of Forensic Sciences

View full text Add to dashboard Cite

Ground truth data (GTD) is used by those in the field of digital forensics (DF) for a variety of purposes including to evaluate the functionality of undocumented, new, or emerging technology and services and the digital traces left behind following their usage. Most accepted and reliable trace interpretations must be derived from an examination of relevant GTD, yet despite the importance of it to the DF community, there is little formal guidance available for supporting those who create it, to do so in a way that ensures any data is of good quality, reliable, and therefore usable. In an attempt to address this issue, this work proposes a minimum standard of documentation that must accompany the production of any GTD, particularly when it is intended for use in the process of discovering new knowledge, proposing original interpretations of a digital trace, or determining the functionality of any technology or service. A template structure is discussed and provided in Appendix S1 which sets out a minimum standard for metadata describing any GTD's production process and content. It is suggested that such an approach can support the maintenance of trust in any GTD and improve the shareability of it.

show abstract

“…Ground-truth originally refers to images or measurements identified by maps, air photography or satellites that can be physically confirmed by on-ground observation. 32,33 It is now a common AI term referring to data accepted to be true, by direct observation or measurement. This provides a reference standard to test the prediction performance of a machine or human observer.…”

Section: Algorithm Trainingmentioning

confidence: 99%

A systematic review of automated segmentation of 3D computed‐tomography scans for volumetric body composition analysis

Van

Drami

Pring

et al. 2023

J cachexia sarcopenia muscle

View full text Add to dashboard Cite

Automated computed tomography (CT) scan segmentation (labelling of pixels according to tissue type) is now possible. This technique is being adapted to achieve three‐dimensional (3D) segmentation of CT scans, opposed to single L3‐slice alone. This systematic review evaluates feasibility and accuracy of automated segmentation of 3D CT scans for volumetric body composition (BC) analysis, as well as current limitations and pitfalls clinicians and researchers should be aware of. OVID Medline, Embase and grey literature databases up to October 2021 were searched. Original studies investigating automated skeletal muscle, visceral and subcutaneous AT segmentation from CT were included. Seven of the 92 studies met inclusion criteria. Variation existed in expertise and numbers of humans performing ground‐truth segmentations used to train algorithms. There was heterogeneity in patient characteristics, pathology and CT phases that segmentation algorithms were developed upon. Reporting of anatomical CT coverage varied, with confusing terminology. Six studies covered volumetric regional slabs rather than the whole body. One study stated the use of whole‐body CT, but it was not clear whether this truly meant head‐to‐fingertip‐to‐toe. Two studies used conventional computer algorithms. The latter five used deep learning (DL), an artificial intelligence technique where algorithms are similarly organized to brain neuronal pathways. Six of seven reported excellent segmentation performance (Dice similarity coefficients > 0.9 per tissue). Internal testing on unseen scans was performed for only four of seven algorithms, whilst only three were tested externally. Trained DL algorithms achieved full CT segmentation in 12 to 75 s versus 25 min for non‐DL techniques. DL enables opportunistic, rapid and automated volumetric BC analysis of CT performed for clinical indications. However, most CT scans do not cover head‐to‐fingertip‐to‐toe; further research must validate using common CT regions to estimate true whole‐body BC, with direct comparison to single lumbar slice. Due to successes of DL, we expect progressive numbers of algorithms to materialize in addition to the seven discussed in this paper. Researchers and clinicians in the field of BC must therefore be aware of pitfalls. High Dice similarity coefficients do not inform the degree to which BC tissues may be under‐ or overestimated and nor does it inform on algorithm precision. Consensus is needed to define accuracy and precision standards for ground‐truth labelling. Creation of a large international, multicentre common CT dataset with BC ground‐truth labels from multiple experts could be a robust solution.

show abstract

On ‘ground’ truth and why we should abandon the term

Cited by 8 publications

References 23 publications

Large disagreements in estimates of urban land across scales and their implications

Large disagreements in estimates of urban land across scales and their implications

A template for creating and sharing ground truth data in digital forensics

A systematic review of automated segmentation of 3D computed‐tomography scans for volumetric body composition analysis

Contact Info

Product

Resources

About