• Home
  • Biometrics
  • Emerging Biometric Technology: Automated Scar, Mark, and Tattoo Identification

Emerging Biometric Technology: Automated Scar, Mark, and Tattoo Identification

According to the American Academy of Dermatology, over 36% of Americans aged 18 to 29 have at least one tattoo, and recently, law enforcement agencies have increased the use of scars, marks and tattoos (SMT) for identification purposes. While several people might have the same tattoo or similar scars and marks, when layered with another biometric such as facial recognition, an automated SMT detection algorithm could increase the overall accuracy of the biometric search.

Current methods for SMT search and identification have limited accuracy

Traditionally, law enforcement agencies perform forensic searches of biometric systems for an unknown person’s SMT by searching written descriptions which are located in pre-defined categories in a database. However, as tattoos have grown in complexity, it has become harder to accurately describe and categorize all variations of a tattoo into one category in the defined standard. For example a tattoo could have both a skull and a flag, however according to current processing standards only one category—skull or flag—could be chosen. Researchers are looking into automated image searching and matching to help make SMT identification more accurate and useful.

Automated image based tattoo matching identifies unique markings

Image of feature points assigned to points of interest on a tattoo image.
Feature points assigned to a tattoo image. Source: Jung-Eun Lee, Anil K. Jain, and Rong Jin

Recent research at Michigan State University has produced a technique that conducts an automated search on tattoo images as opposed to manual textual searches. This technique focuses on extracting repeatable characteristic feature points from a tattoo picture for matching, and it applies filters to remove distortions and blurriness that can occur in field captures.

The algorithm used is known as a Scale-Invariant Feature Transform (SIFT). The algorithm extracts feature points from a tattoo image, and then matching algorithms perform a distance comparison of feature points to find a match. The researchers at Michigan State found this method to be unaffected by spatial distortion and image rotation.

Automated processes detect and classify scars, marks, and tattoos

Image based tattoo matching has been promising when dealing with images taken in a controlled environment, such as by law enforcement officials during the booking process. However, a vast source of tattoo images linked to identities is publicly available on the internet. Recent research at the University of Colorado-Colorado Springs provides a method for identifying tattoos from uncontrolled images, or those found “in the wild” using a graph-based visual saliency algorithm (GBVS). The image below illustrates the use of this algorithm on a publicly accessible internet image. The GBVS predicts which areas of an image will attract the attention of a viewer in that these areas stand out in some way from the background. When this algorithm is teamed with an automated segmentation tool, users can extract tattoo images in an automated process.

Image of GBVS on a typical publically accessible internet image.
GBVS on a typical publicaly accessible internet image. Brian Heflin, Walter Scheirer, T.E. Boult. Detecting and Classifying Scars, Marks, and Tattoos Found in the Wild.

The use of scars, marks, and tattoos for biometric identification could complement and enhance methods currently in place

SMT image based matching could provide more accurate results when identifying a subject by returning additional information to supplement a traditional fingerprint search. In addition, technology that detects and classifies SMT in the wild could be used to build and maintain a database of exemplar gang tattoos from publicly accessible photographs. This information could provide officers with additional context when identifying a subject.


About the Author: Mike Wagner is an IEEE Certified Biometric Professional and has performed biometrics development on projects such as US-VISIT and the system used by the Border Patrol to collect biometric data in the field. In addition to his focus in biometrics, Mike Wagner also specializes in enterprise architecture, Java software development, large system troubleshooting and problem solving.