NIST workshop discusses new approaches to improving tattoo recognition
An international group of experts from industry, academia and government participated in National Institute of Standards and Technology’s (NIST) Tattoo Recognition Technology Challenge Workshop in which it discussed challenges and potential ways of automating tattoo recognition to help law enforcement identify criminals and victims.
The move comes a few months after the National Institute of Standards and Technology invited all interested commercial and academic organizations to participate in advancing research and development of automated image-based tattoo matching technology.
After being initially presented with the results of a preliminary trial of existing tattoo recognition software, industry and academia were asked by the FBI Biometric Center of Excellence (BCOE) to take initial steps into automated image-based tattoo matching technology.
The current method of cataloging tattoo images for the purpose of sharing among various law enforcement agencies is completely reliant on a keyword based process.
However, multiple keywords must be used in a database to keep up with the increasing selection of tattoo designs requires multiple keywords, which can result in the same tattoo being given different labels depending on the examiner.
All of the participating organizations used the same BCOE-provided dataset of thousands of images from government databases.
NIST offered participants five use cases and asked them to report their performance on finding
visually similar or related tattoos from different subjects, different instances of the same tattoo image from the same subject over time, a small region of interest that is contained in a larger image, visually similar or related tattoos using different types of images, and whether an image contains a tattoo or not.
“The state-of-the-art algorithms fared quite well in detecting tattoos, finding different instances of the same tattoo from the same subject over time, and finding a small part of a tattoo within a larger tattoo,” said NIST computer scientist Mei Ngan, who organized the challenge.
Meanwhile, Ngan discovered that were two areas that could use further research, including detecting visually similar tattoos on different people and recognizing a tattoo image from a sketch or sources other than a photo.
“Improving the quality of tattoo images during collection is another area that may also improve recognition accuracy,” said Ngan.
There are several organizations that participated in the challenge including Compass Technical Consulting, LLC., the Fraunhofer Institute of Optronics, System Technologies and Image Exploitation, the French Alternative Energies and Atomic Energy Commission, MITRE, MorphoTrak and Purdue University.
Workshop participants also explored the potential use of image-based tattoo matching in operations, identified notable discrepancies and the need to improve tattoo recognition, as well as the next steps NIST might take in this area.