Commons:How to detect copyright violations
How to detect copyright violations
[edit]Often, people upload images, particularly photographs, that do not abide by our licensing requirements. This is done sometimes in good faith, sometimes in bad faith. So here a brief how-to for detecting many copyvios with a high probability.
It is not possible for us to detect all copyright violations. However, we can detect a great number of them by simply having a look around, especially in Special:Newimages, and watching for tell-tale signs. Here is a list of signs of probable copyright violations:
- Photographs by professional photographers
- Few of these grant free licenses over their images. Consider contacting the photographer, if possible using the e-mail templates, if somebody posts a vague claim to be the photographer, or of having obtained authorization.
- In the event that a professional photographer allows reproduction of their images, make sure that the images are licensed under a free license (i.e. not "noncommercial" or "Wikipedia only") and post the permission to permissions-commons
wikimedia.org.
- Photographs of celebrities, rock bands, etc.
- Not all such photos are copyright violations, but they often are, and thus they warrant special scrutiny. Signs of probable violations:
- "Staged" photographs, where the subjects obviously pose for the photographer. Almost always, such people do not pose for amateur photographers; they pose for (semi-)professionals, who usually demand payment for their works and seldom put them under free licenses.
- Low-resolution pictures. Probably taken from a web site; photographers have access to better content.
- It is however possible for amateurs to take photographs of celebrities. Examples include concerts, public appearances, etc.; this often needs chance. But, generally, the people will then explain how and where they took the photograph, and will provide some high-resolution shot.
Watermarks
- If the photograph contains a watermark, chances are it was not released under a free license.
- Photographs of items or from angles not accessible to the general public
- Generally, only professional crews have access to "press only" locations offering a good vantage point on the starting lines of races, etc.
- Thumbnails (very low resolution photographs)
- These are a tell-tale sign that the image has been downloaded from a web site.
- File names
- Servers often generate special file names.
- facebook: ddddd_dddddddddddd_dddddddd_ddddddd_ddddddd_D.jpg → http://www.facebook.com/profile.php?id=dddddddd
- flickr: http://server.staticflickr.com/dddd/dddddddddd_hexhexhex_size.jpg → getInfo → (photopage)
- Meta data
- Especially the software, copyright, author, make and model rows are interesting. If you don't know the software, look it up. If it is a server-software, you are alerted.
- Suspicious licenses
- Apart from screenshots and icons of free software computer programs, very few images are licensed under the GNU GPL or other free software license. As well many people randomly tag images with "PD-Self" and such. Some people apparently choose a random free license to upload copyrighted content.
- Modern art
- Paintings, sculptures, and other works of art are copyrighted by their author (at least, in most jurisdictions). Photographs of such items can thus be covered by copyright of the artist (depending on the jurisdiction and the location of the item). Check for acceptable exceptions from this rule: {{PD-Art}} and {{FoP}}.
- Suspicious user
- Users who repeatedly upload copyright protected items under false licenses. These users may have several warnings, and many attempt to hide the original copyright owners identity by stating they created the work. Questionable Flickr images lists users from Flickr, the popular image hosting site, that we believe have incorrectly marked photographs as free that they don't have the right to.
Search engines
[edit]Search engines may help you to find the original source of an image.
For Google search-by-image and Tineye, there is a gadget in your preferences (Maintenance tools) adding a one-click link. Alternatively, you can use a browser addon like the TinEye Reverse Image Search addon for Firefox.
https://tineye.com is a "reverse image search engine" – you enter the URL of the image, and it searches for the same image. While it has only indexed a small fraction of the images on the Internet, it's quite good at identifying stock photographs from large providers. Tineye also reports the dates it indexed individual images, which makes it easier to evaluate whether an image uploaded to Commons is actually someone's original work.
Other than Tineye, the largest 4 image search engines are:
- https://images.google.com/ (sometimes it works much better than Tineye)
- https://www.google.com/intl/es419/insidesearch/features/images/searchbyimage.html (searching for exact and similar images)
- https://yandex.kz/images/
- http://images.duckduckgo.com/
- https://images.search.yahoo.com/
- https://www.bing.com/?scope=images
See also
[edit]- About copyright violations
- Previously deleted files
- Commons:Abuse filter/Automated copyvio detection
- Proposal for bots detecting copyvios on Commons
Categories to scan
[edit]- Sub-categories of media needing categories
- Mobile uploads lacking EXIF data for camera
- Videos featuring unidentified music
- One can also scan the New files gallery or searches sorted by recency e.g. for new audio files, possibly using a bot tagging likely copyvios