HRT Transgender Dataset
Report incident ๐ฅ | Improve page ๐ | Access database ๐ข
Created in 2013, the HRT Transgender Dataset helps facial recognition systems to identify users of Hormone Replacement Therapy (HRT) transitioning or who have already transitioned from one gender to another.ย
The dataset comprised 10,000 images of 38 people, with an average of 278 images per subject taken from publicly available YouTube videos 'under real-world conditions, including variations in pose, illumination, expression, and occlusion'.ย
Dataset ๐ค
Removed
Operator:ย
Developer: University of North Carolina, Wilmington (UNCW)
Country: USA
Sector: Research/academia; Technology
Purpose: Identify HRT users
Technology: Database/dataset; Facial recognition; Computer vision
Issue: Copyright; Ethics/values; Bias/discrimination - LGBTQ; Privacy
Transparency: Governance; Marketing; Privacy
Risks and harms ๐
The HRT Transgender dataset has been criticised for using and exposing the data of YouTube users without consent, potential bias against the LGBTQ community, and poor transparency.
Transparency and accountability ๐
The HRT Transgender dataset is seen to have multiple transparency limitations:
Limited access to raw data. Researchers attempting to audit the database have faced difficulties in accessing the original, unprocessed data.
Lack of clear documentation. There appears to be insufficient documentation regarding the data collection methods, inclusion criteria, and data processing procedures used to compile the database.
Potential selection bias. The database may not be representative of the entire transgender population, as it likely only includes individuals who have disclosed their transgender identity to healthcare providers and had it recorded in their medical records.
Incidents and issues ๐ฅ
Research, advocacy ๐งฎ
Keyes, O., Austin, J. (2022). Feeling fixes: Mess and emotion in algorithmic audits
Scheuerman M.K., Pape M., Hanna A. (2021). Auto-essentialization: Gender in automated facial analysis as extended colonial project (pdf)
Kumar V., Ramachandra R., Namboodiri A.M., Busch C. (2016). Robust transgender face recognition: Approach based on appearance and therapy factors
Page info
Type: Data
Published: January 2023
Last updated: June 2024