Coronavirus Mask Image dataset
The Coronavirus Mask Image Dataset is a dataset of 1,200 people wearing face masks in order to improve the quality of biometric facial recognition algorithms . The dataset was created during the COVID-19 pandemic by US-based AI startup WorkAround . The company was criticised for using people's photographs posted to Instagram without acknowledgement or consent, thereby violating Meta's usage policies and violating peoples' privacy. Dataset 🤖 Data 🔗 Released: 2020 Developer: WorkAround Sector: Health Purpose: Train facial recognition systems Type: Database/dat aset Technique: Computer vision; Facial recognition Transparency, accountability 🙈 The Coronavirus Mask Image Dataset is seen to suffer from several important transparency limitations, including: Data collection methods. The exact methods used to gather the images are not fully disclosed. Source diversity. It is unclear how diverse the sources of the images are, which could lead to biases. Privacy consent. It is unclear whether individuals in the images consented to their inclusion. Annotation process. Details about how images were labeled (e.g., "mask worn correctly," "no mask") are limited. Dataset composition. The balance between masked and unmasked images, or different types of masks, is not well-documented. Demographic representation. Information about the diversity of individuals represented in the dataset is lacking. Image quality variation. The range of image qualities and how this might affect model performance is not described. Context information. Details about where and when images were taken (e.g., indoor/outdoor, pre/post-pandemic) are missing. Risks, harms 🛑 The Coronavirus Mask Image dataset has been cri ticised for violating the privacy of the general public in an inappropriate and unethical manner. Incidents, issues 🔥 Coronavirus Mask Image dataset criticised for violating user privacy
- Organisation involved
- WorkAround
- Product, system or model
- Coronavirus Mask Image Dataset
This incident was imported from AIAAIC and is used under CC BY-SA 4.0. Our additions to it — the structured fields, the translation, the checks against other reports — are published under the same licence.
This is a record of what was reported, not a finding that anyone broke the law. If it names your organisation and you believe it is wrong, the corrections process is free and open to everyone.