GUIDE-AI: Guidelines for Unified Imaging Dataset Development and Assessment for Artificial Intelligence
Fuente:
Zenodo
Saved in:
| Main Author: | |
|---|---|
| Format: | Recurso digital |
| Language: | English |
| Published: |
Zenodo
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866902307831021568 |
|---|---|
| author | Pinto-Coelho, Luís |
| author_facet | Pinto-Coelho, Luís |
| contents | <p>The integrity and methodological rigor of medical imaging datasets have become foundational prerequisites for clinically translatable, safe AI systems. Despite this, most existing datasets are assembled through opportunistic, unstructured approaches that prioritize data availability over scientific soundness, resulting in annotation inconsistencies, selection biases, and demographic underrepresentation. These structural weaknesses are further compounded by the rapid proliferation of generative data synthesis, foundation model dependencies, and increasingly demanding regulatory instruments, notably the EU AI Act. We introduce GUIDE-AI, a principled operational framework for the unified development and rigorous assessment of medical imaging datasets. The framework integrates two mutually reinforcing components: a governance pipeline ensuring ethical oversight, legal compliance, accountability, and stakeholder stewardship; and a technical pipeline translating these principles into reproducible engineering workflows across problem formulation, data acquisition, annotation protocols, quality control, and curated dataset release. By treating governance as a first-class design requirement throughout the dataset lifecycle, GUIDE-AI operationalizes trustworthy AI beyond rhetorical commitment, offering the field a concrete, scalable pathway toward datasets that are scientifically rigorous, institutionally transparent, and regulation-ready. </p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_19916897 |
| institution | Zenodo |
| language | eng |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | GUIDE-AI: Guidelines for Unified Imaging Dataset Development and Assessment for Artificial Intelligence Pinto-Coelho, Luís Artificial Intelligence Diagnostic Imaging Machine Learning Databases, Factual Data Curation Information Management <p>The integrity and methodological rigor of medical imaging datasets have become foundational prerequisites for clinically translatable, safe AI systems. Despite this, most existing datasets are assembled through opportunistic, unstructured approaches that prioritize data availability over scientific soundness, resulting in annotation inconsistencies, selection biases, and demographic underrepresentation. These structural weaknesses are further compounded by the rapid proliferation of generative data synthesis, foundation model dependencies, and increasingly demanding regulatory instruments, notably the EU AI Act. We introduce GUIDE-AI, a principled operational framework for the unified development and rigorous assessment of medical imaging datasets. The framework integrates two mutually reinforcing components: a governance pipeline ensuring ethical oversight, legal compliance, accountability, and stakeholder stewardship; and a technical pipeline translating these principles into reproducible engineering workflows across problem formulation, data acquisition, annotation protocols, quality control, and curated dataset release. By treating governance as a first-class design requirement throughout the dataset lifecycle, GUIDE-AI operationalizes trustworthy AI beyond rhetorical commitment, offering the field a concrete, scalable pathway toward datasets that are scientifically rigorous, institutionally transparent, and regulation-ready. </p> |
| title | GUIDE-AI: Guidelines for Unified Imaging Dataset Development and Assessment for Artificial Intelligence |
| topic | Artificial Intelligence Diagnostic Imaging Machine Learning Databases, Factual Data Curation Information Management |
| url | https://doi.org/10.5281/zenodo.19916897 |