Law enforcement agencies and governments have long used biometric modalities for accurate identification, enabling law enforcement activities and immigration and border controls.
Our data can be collected and lawfully used through facial, iris or fingerprint recognition tools and, now, a new biometric modality has emerged in tattoo recognition technology, which can verify who someone is by the permanent visible marking on their body.
It may not be part of an individual’s DNA or genetic makeup, however a tattoo modifies the body and can contribute to identifying unique features about someone to help authorities in investigations.
We have introduced a panel at Identity Week America on NIST’s world-class evaluation of tattoo recognition, titled “Tatt-E: Benchmarking the state of the art in tattoo recognition”. Find the session on our agenda.
Historically, authorities have searched tattoo image databases using text descriptions of tattoos to assist in their investigations. However, the effectiveness of this method has been limited by the subjectivity of text-based descriptions. Recent advances in technology have enabled developers to leverage artificial intelligence (AI) to create automated, image-based tattoo search capabilities.
Compared to traditional approaches that rely on subjective text descriptions, image-based searching provides a more objective means of retrieval. To determine whether these systems are fit for purpose, decision-makers will need to know their capabilities and limitations. NIST is running an evaluation program to assess the accuracy of tattoo recognition algorithms.
This evaluation will measure the capability of these algorithms to detect tattoos in an image and to perform automated matching of different images of the same tattoo from the same subject over time. This talk, presented by Mei Ngan, a scientist at NIST, will present the state-of-the-art accuracy in image-based tattoo recognition and discuss current capabilities and limitations of the technology.
NIST evaluates tattoo recognition technology as a secondary biometric modality designed to assist law enforcement. While traditional biometric standards classify tattoos by visual content categories (ANSI/NIST-ITL), NIST’s evaluation programs categorise the technology itself by its capability to detect, locate, match instances, match across media types, and measure visual similarity.
Why not watch this session…
Face recognition: seven areas for further gain
With Patrick Grother, Lead Scientist, NIST
Face recognition has capability has improved massively over the last decade, benefitting from AI research in machine learning, deep neural networks, and computer vision. New capability has enabled more developers to offer an expanded and broad array of applications. Yet some problems remain, and the talk will describe seven areas where further technical advancements are needed. Additionally, the presentation will outline how current standardisation activities and the recent expansion of NIST’s face recognition evaluations could improve capabilities in some of these areas.
