Face Forensics  Image Recognition Suite
Advanced Face Recognition           Tattoo Recognition
Partial Face Recognition                 Scene Recognition

    Tattoo Recognition

f2 Tattoo is an independent but fully integrated module within the f2 Image Recognition Suite.

As with f2 face recognition, tattoos are first enrolled by determining and encoding a number of specific characteristics of the image. The nature and relative position of these is transformed into a numeric string, which is stored as an encode array together with the record ID in f2’s own internal database. This process is complex and will typically take a few seconds for each image. The resulting encode arrays are held in memory on a matching server, enabling subsequent searches to be extremely fast.  f2 analyzes images in read-only mode and stores the resulting non-reversible encode arrays – it doesn’t store the images themselves.

The f2 Tattoo Module is designed to match an unknown tattoo against a database of tattoos in order to determine if there is a matching tattoo (or tattoos) there. Unlike hashing algorithms, f2 will match a cropped, re-saved, or reformatted tattoo against those in the database. Results are displayed as thumbnails of potentially matching tattoos, in descending order of Match%. Any of these can be selected to display the face and name of the owner in the original database record.

The result of a Tattoo search will be a number of thumbnail images that have the same or similar characteristics to the image being searched (the “probe” image), above a user-defined threshold. An officer can then visually scan the images to determine which, if any, is the same tattoo. The face associated with any selected tattoo is displayed, together with a link back to the original record. The benefit of f2 is therefore that it can find a matching tattoo in seconds where an officer could take many hours or longer to find a match.
f2 Tattoo is not intended to identify the same type of tattoo, e.g. an insect such as a scorpion, if the images are significantly different. For example it is unlikely to match a stylized drawing of a scorpion against a photograph. In general, at least in a law enforcement environment, such searches can be narrowed down by searching the existing database against a text string such as “scorpion”. f2 will display potential matches within the group that meets this description.