Commons:Photographs of identifiable people/Update 2013/Identification

Scope Review 2013 links: Discuss stage 2 of this review Translation Help your community by translating the review Background Main review page: Background and process FAQ Links to current rules Project Scope Photographs of identifiable people Discussion: Introductory Scope wording Aim of Wikimedia Commons Excluded educational content Discussion: Files Must be a media file Must be an allowable free file format Must be freely licensed or public domain Licensing terms Evidence Precautionary principle 'Realistically useful for an educational purpose' PDF and DjVu formats Discussion: Pages, galleries and categories General pages, galleries and categories User pages, galleries and categories Discussion: Areas of particular concern Censorship "Neutral point of view" Discussion: Identifiable people Lead Legal issues Consent Moral issues Identification Examples Removal requests Other image sources Other proposals Other proposals

Identification

The degree to which a subject is identifiable varies. An image that includes a clear view of the face is highly identifiable. Other features of the person's body, clothing or the location may help with identifying the subject. Outside of the image, clues may be obtained from the image title, description, origin, source url, geolocation and date. The greater the privacy issue with an image, the more weight should be given to the risk of identification by non-obvious means. Whether the person is the clear subject of the photograph or a mere bystander or background detail is another important factor.

The risk of identification can be minimised by not including certain information in the image description. However some details regarding the origin of the image (such as source url and author) may be a requirement of the source image licence or Commons policy, so cannot be removed. It may also be possible to shoot the subject from a different angle or frame it differently.

It is better to obtain consent than to attempt to anonymise an image that may require it. Placing a black band over the eyes was historically used to hide patient identity in medical publications but is no longer considered effective.(Source: ICMJE) Pixelated features can sometimes be revealed by squinting one's eyes. These crude attempts to anonymise images may damage the value of an image to Commons to such a degree that it has limited or no realistic chance of being used.

Where the law forbids taking or publishing a photograph of a person without consent, and consent has not been given, then making the subject hard to identify (such as blurring their face) is unethical: the photograph should not be uploaded to Commons.

If the original or similar images are already present on the internet (either on Commons or elsewhere) then attempts at anonymising the subject are ultimately futile. Content-based image retrieval engines such as TinEye or Google Images can identify a subject that has been anonymised. All of the following people are readily identifiable by anyone familiar with the subject. They may also be identified by computer, by simply dragging and dropping the image onto Google Images.

• Problems with attempts at anonymisation
• 
  Black band over the eyes
• 
  Pixelated face
• 
  Cropped head